Design and Implementation of a Hybrid Technology ...
Transcript of Design and Implementation of a Hybrid Technology ...
University of Nebraska - Lincoln University of Nebraska - Lincoln
DigitalCommons@University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln
Theses, Dissertations, & Student Research in Computer Electronics & Engineering
Electrical & Computer Engineering, Department of
Spring 4-18-2013
Design and Implementation of a Hybrid Technology Networking Design and Implementation of a Hybrid Technology Networking
System System
Sushanta Mohan Rakshit University of Nebraska-Lincoln, [email protected]
Follow this and additional works at: https://digitalcommons.unl.edu/ceendiss
Part of the Digital Communications and Networking Commons
Rakshit, Sushanta Mohan, "Design and Implementation of a Hybrid Technology Networking System" (2013). Theses, Dissertations, & Student Research in Computer Electronics & Engineering. 22. https://digitalcommons.unl.edu/ceendiss/22
This Article is brought to you for free and open access by the Electrical & Computer Engineering, Department of at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Theses, Dissertations, & Student Research in Computer Electronics & Engineering by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln.
DESIGN AND IMPLEMENTATION OF A HYBRID TECHNOLOGY
NETWORKING SYSTEM
By
Sushanta Mohan Rakshit
A THESIS
Presented to the Faculty of
The Graduate College at the University of Nebraska
In Partial Fulfillment of Requirements
For the Degree of Master of Science
Major: Telecommunications Engineering
Under the Supervision of Professor Hamid R. Sharif-Kashani
Lincoln, Nebraska
May 2013
DESIGN AND IMPLEMENTATION OF A HYBRID TECHNOLOGY
NETWORKING SYSTEM
Sushanta Mohan Rakshit, M.S.
University of Nebraska, 2013
Advisor: Hamid R. Sharif-Kashani
The safety of rail transport has always been the top priority for the Federal
Railroad Administration (FRA). Legacy technology, like wayside monitoring, is still
in place and is largely relied upon for detection of faults. Modern technology like
Radio Frequency Identification (RFID) has been introduced recently. However, this
is largely used to detect a particular railcar rather than to monitor it for problems.
Wireless Sensor Network (WSN) technology is being evaluated by the railroads for
real-time or near real-time monitoring of the status of railcars for timely response to
problems and also for trend analysis.
ZigBee has been the networking protocol of choice for the railroads for its low
power consumption and cost of implementation. The railroad scenario presents a long
linear-chain like network topology which ZigBee was not designed to handle. It has
been found that a ZigBee-only network in the railroad environment suffers from
drawbacks like long synchronization delays, severe problems with route discovery
and maintenance, aggregation of data errors leading to unacceptable packet loss rates,
lack of a mechanism to decide traffic priority for critical packets, like alarm, so that
they can reliably traverse the network to the collecting node in the locomotive etc.
Hybrid Technology Networking (HTN) protocol has been suggested which
addresses the shortcomings of ZigBee in the railroad scenario. It proposes a
standards-based multi-protocol approach that is well-suited for the railroad scenario.
The current crop of sensor platforms does not provide an integrated environment for
the implementation of HTN.
In this research work an integrated hardware platform for the implementation
of the HTN protocol is designed and implemented. The guiding principle has been the
adherence to standards. The test results using the hardware show that it provides inter-
operability with available sensor platforms, can interface with other sensing hardware
using standard protocols and provides communication capabilities exceeding that
needed by HTN.
v
Dedicated to the countless thinkers and innovators who push the boundaries of
knowledge today, for a better tomorrow.
vi
Acknowledgments
I would, first and foremost, like to thank my advisor Dr. Sharif for his
invaluable guidance and support during the course of my research work. I would also
like to thank Dr. Michael Hempel for his timely inputs and new perspectives on the
work that I was doing. I am also thankful to my colleagues Fahimeh, Pradhumna and
Tao for their help and support during the time I was working on this research. I would
specially like to mention Fahimeh for her unwavering support in helping me with
laying out the PCB and teaching me how to work with the Cadence software suite.
I would like to sincerely thank the Federal Railroad Administration (FRA) for
their generous funding of the project that has led to this research. I would like to
mention John Punwani and Monique Stewart for their unwavering support of the
project.
Although no amount of gratitude can ever be enough for them but still I would
like to thank my parents for their unwavering support and belief in my dreams. I
would not be here today writing these words without their constant inspiration and
encouragement. Last, but not in the very least, a special mention for the angel with a
flower tiara, who has looked over me and guided me through it all.
vii
Table of Contents
List of Figures …………………………………………………………………….. xi
List of Tables ……………………………………………………………………... xv
List of Acronyms ………………………………………………………………..... xvi
Chapter 1. INTRODUCTION TO STATUS MONITORING IN RAILROADS … 1
1.1.Importance of Freight Railroad ……………………………………………… 1
1.2.Reasons for Monitoring Freight Railroad …………………………………… 3
1.3.Existing Methods for Railcar Status Monitoring …………………………… 4
1.4.Drawbacks of Current Methods …………………………………………….. 6
1.5.WSN for Railcar Status Monitoring ………………………………………… 7
Chapter 2. OVERVIEW OF COMMUNICATION PROTOCOLS ……………… 8
2.1.IEEE 802.11 ………………………………………………………………… 8
2.1.1.IEEE 802.11 Components ………………………………………………. 9
2.1.2.IEEE 802.11 Network Types …………………………………………… 10
2.1.3.IEEE 802.11 Protocol Stack …………………………………………….. 12
2.1.3.1.IEEE 802.11 Physical Layer ………………………………………… 12
2.1.3.2.IEEE 802.11 Data Link Layer ………………………………………. 14
2.1.4.IEEE 802.11 Security …………………………………………………… 15
2.2. IEEE 802.15.4 ……………………………………………………………… 15
2.2.1.LR-WPAN Device Architecture ………………………………………… 16
2.2.2.IEEE 802.15.4 Physical Layer ………………………………………….. 17
2.2.3.IEEE 802.15.4 MAC Layer ……………………………………………... 20
2.2.4.IEEE 802.15.4 Network Topologies ……………………………………. 25
2.2.5.IEEE 802.15.4 Device Types ………………………………………........ 26
viii
2.3.ZigBee ………………………………………………………………………. 27
2.3.1.ZigBee Stack Architecture ……………………………………………... 27
2.3.1.1.ZigBee Network Layer (NWK) ……………………………………... 28
2.3.1.2.ZigBee Application Layer …………………………………………... 31
2.3.2.ZigBee Device Types ………………………………………………........ 32
2.3.3.ZigBee Network Topologies ……………………………………………. 33
Chapter 3. ISSUES WITH ZIGBEE IN RAILROAD ENVIRONMENTS ……… 34
3.1. Route Discovery ……………………………………………………………. 35
3.2.Synchronization Delay ……………………………………………………… 35
3.3.Packet and Link Loss ……………………………………………………….. 36
3.4.Lack of Quality-of-Service …………………………………………………. 37
3.5.Data Forwarding and Aggregation …………………………………………. 38
3.6.Network Lifetime …………………………………………………………... 39
Chapter 4. OVERVIEW OF HYBRID TECHNOLOGY NETWORKING …….. 41
Chapter 5. PROBLEM STATEMENT …………………………………………… 44
Chapter 6. HARDWARE DESIGN ……………………………………………… 46
6.1.Requirements ……………………………………………………………….. 46
6.2.Block Diagram of Design ………………………………………………...... 48
6.3.Hardware Components …………………………………………………….. 49
6.3.1.Power Supply Circuitry ………………………………………………… 49
6.3.2.Clock Circuitry ………………………………………………………… 49
6.3.3.Microcontroller ………………………………………………………… 50
6.3.4.Wi-Fi Module ………………………………………………………….. 51
6.3.5.ZigBee Module ………………………………………………………… 53
ix
6.3.6.SDHC Module ………………………………………………………….. 54
6.3.7.USB Module ……………………………………………………………. 55
6.3.8.RS-232 Module …………………………………………………………. 55
6.3.9.LCD Module ……………………………………………………………. 56
6.3.10.External Sensor Board Expansion Slot ……………………………….. 58
6.3.11.PCB Layou t …………………………………………………………… 58
6.3.12.HTNMote in Pictures ………………………………………………..... 60
Chapter 7. SOFTWARE IMPLEMENTATION …………………………………. 66
7.1.Wi-Fi Implementation ……………………………………………………… 66
7.2.ZigBee Implementation …………………………………………………….. 71
7.3.SDHC Implementation ……………………………………………………... 74
7.4.USB Implementation ……………………………………………………….. 74
7.5.RS-232 Implementation …………………………………………………….. 74
7.6.LCD Implementation ……………………………………………………….. 75
Chapter 8. TESTS AND TEST RESULT ANALYSIS …………………………… 77
8.1.Performance Tests …………………………………………………………… 77
8.1.1.Methodology ……………………………………………………………. 77
8.1.1.1. Current Consumption Test Setup …………………………………… 78
8.1.1.2. Channel Emulator ………………………………………………….. 78
8.1.1.3. Wi-Fi Tests Setup ………………………………………………….. 80
8.1.1.4. ZigBee Tests Setup ……………………………………………….... 83
8.1.2. Current Consumption Test Results ……………………………………... 85
8.1.3. Wi-Fi Test Results ……………………………………………………… 87
8.1.4. ZigBee Test Results ……………………………………………………. 91
x
8.1.5. Tests Result Analysis ………………………………………………….. 94
8.1.5.1. Wi-Fi Results Analysis …………………………………………….. 94
8.1.5.2. ZigBee Results Analysis …………………………………………… 95
8.2. Functionality Tests …………………………………………………………. 95
8.2.1. USB Test ……………………………………………………………….. 96
8.2.2. RS-232 Test …………………………………………………………….. 97
8.2.3. LCD Test ……………………………………………………………….. 98
8.2.4. SDHC Test ……………………………………………………………… 99
Chapter 9. SUMMARY, CONCLUSION AND FUTURE WORK ……………… 100
9.1.Summary and Conclusion …………………………………………………… 100
9.2.Future Work …………………………………………………………………. 102
REFERENCES ……………………………………………………………………. 103
APPENDIX A …………………………………………………………………….. 109
APPENDIX B …………………………………………………………………….. 109
APPENDIX C …………………………………………………………………….. 109
xi
List of Figures
Figure 1.1: Percent in ton-miles of freight carried by different
modes of transport …………………………………………………... 1
Figure 1.2: 2005 Class I Railroads Revenues by Commodity Group ……………. 2
Figure 1.3: Wayside monitoring equipment in the field …………………………. 4
Figure 1.4: A railcar with an AEI tag mounted on its side ………………………. 5
Figure 2.1: The IEEE 802.11 Infrastructure based network architecture ………... 10
Figure 2.2: A few laptops in an ad-hoc IEEE 802.11 network ………………….. 11
Figure 2.3: IEEE 802.11 Protocol Stack ………………………………………… 12
Figure 2.4: Channel assignment in the 2.4 GHz band for IEEE 802.11 ………… 13
Figure 2.5: LR-WPAN Device Architecture ……………………………………. 17
Figure 2.6: Format of a IEEE 802.15.4 PPDU ………………………………….. 19
Figure 2.7: IEEE 802.15.4 MAC Frame format ………………………………… 21
Figure 2.8: IEEE 802.l5.4 MAC Frame control field …………………………… 22
Figure 2.9: Sequence of steps when Coordinator has
data to send to Network device in a beacon enabled mode ………... 24
Figure 2.10: Sequence of steps when the Network device has
data for Coordinator in a beacon enabled mode …………………. 24
Figure 2.11: Sequence of steps for transfer of data in a
non-beacon enabled network ……………………………………. 25
Figure 2.12: IEEE 802.15.4 network topologies a) Star and b) Peer-to-Peer ….. 26
Figure 2.13: ZigBee stack architecture ………………………………………… 28
Figure 2.14: General ZigBee NWK Layer Frame Format …………………….. 29
Figure 2.15: ZigBee Tree Topology …………………………………………... 33
xii
Figure 3.1: Number of unsuccessful route requests vs. Node ID ………………… 35
Figure 3.2: Time required for synchronization vs. the maximum
number of hops in network ………………………………………….. 36
Figure 3.3: Packet Loss Rate percentage for different hop
counts and Bit Error Rates ………………………………………….. 37
Figure 3.4: Amount of data present in channel as a progression of
hops from the end of train ………………………………………….. 38
Figure 3.5: Energy consumption per node for a 100-railcar train ………………. 39
Figure 3.6: Battery life in hours for each node in a 100-railcar train …………… 40
Figure 4.1: High-Level representation of Hybrid Technology Networking ……. 41
Figure 4.2: Time required for synchronization vs. the maximum
number of hops in network in HTN ………………………………. 42
Figure 4.3: Per Node Throughput vs. the maximum
number of hops in HTN …………………………………………… 42
Figure 4.4: Node Lifetime vs. Total Number of Hops in HTN ………………… 43
Figure 6.1: Hardware Block Diagram ………………………………………….. 48
Figure 6.2: General high level block diagram of the
Redpine Signals Wi-Fi module …………………………………… 52
Figure 6.3: Texas Instruments CC2530 System-on-Chip ………………………. 53
Figure 6.4: Displaytech SD028ATFT LCD with a picture displayed ………….. 57
Figure 6.5: Power Board Top View ……………………………………………. 60
Figure 6.6: Power Board Bottom View …………………………………………. 61
Figure 6.7: Main Board Top View ………………………………………………. 62
Figure 6.8: Main Board Bottom View …………………………………………… 63
xiii
Figure 6.9: HTNMote – Assembled ……………………………………………... 64
Figure 6.10: HTNMote – In action ………………………………………………. 65
Figure 7.1: Slave initialization procedure ………………………………………… 67
Figure 7.2: Memory Read or Master Read procedure ……………………………. 68
Figure 7.3: Frame read or Slave read procedure …………………………………. 69
Figure 7.4: Memory Write or Master Write procedure …………………………... 70
Figure 7.5: Frame write or Slave write procedure ……………………………….. 70
Figure 7.6: Register read procedure ……………………………………………… 71
Figure 7.7: CC2530 in ZNP configuration (Application Processor is K60) ……… 72
Figure 7.8: eGUI block diagram ………………………………………………….. 76
Figure 8.1: ACE 400WB wireless channel emulator ……………………………... 78
Figure 8.2: Power delay profile of the ITU Butler model ………………………… 80
Figure 8.3: Layout and device interconnections for Wi-Fi tests …………………. 81
Figure 8.4: The topology for the Wi-Fi tests ……………………………………... 82
Figure 8.5: Layout and device interconnections for ZigBee tests ………………... 83
Figure 8.6: ZigBee tests channel emulator topology …………………………….. 84
Figure 8.7: Throughput vs. Packet size test result ………………………………... 87
Figure 8.8: Packet loss vs. Path loss result ……………………………………….. 88
Figure 8.9: Packet loss percentage vs. Path Loss result ………………………….. 89
Figure 8.10: ZigBee throughput vs path loss test result ………………………….. 91
Figure 8.11: ZigBee goodput vs path loss test result …………………………….. 92
Figure 8.12: ZigBee packet loss rate vs path loss test results …………………… 93
Figure 8.13: Device Manager screenshot for USB functionality test …………….. 96
Figure 8.14: RS-232 bi-directional communication screenshot ………………….. 97
xiv
Figure 8.15: LCD output after code execution …………………………………… 98
Figure 8.16: Partial screenshot of CodeWarrior console after
SDHC code has executed …………………………………………… 99
xv
List of Tables
Table 2.1: IEEE 802.11 releases over the years …………………………………. 8
Table 2.2: IEEE 802.15.4 versions ………………………………………………. 16
Table 2.3: IEEE 802.15.4 Frequency bands of operation ………………………... 18
Table 8.1: HTNMote Current Consumption ......................................................... 87
xvi
List of Acronyms
ADC Analog to Digital Converter
AEI Automatic Equipment Identification
AF Application Framework
AIB APS Information Base
AP Access Point
API Application Programming Interface
APS Application Support Sub-layer
APSDE APS Data Entity
APSME APS Management Entity
AREQ Asynchronous Request
ARP Address Resolution Protocol
bps Bits per second
CAN Controller Area Network
CRC Cyclic Redundancy Check
CSMA/CA Carrier Sense Multiple Access/Collision
Avoidance
CTS Clear to Send
DAC Digital to Analog Converter
DCF Distributed Coordination Function
DHCP Dynamic Host Configuration Protocol
DSS Distribution System Services
DSSS Direct Sequence Spread Spectrum
ED Energy Detection
xvii
EPA Environmental Protection Agency
FCC Federal Communications Commission
FCF Frame Control Field
FCS Frame Check Sequence
FFD Full Function Device
FHSS Frequency Hopping Spread Spectrum
FRA Federal Railroad Administration
GHz Giga Hertz
GPIO General Purpose Input Output
GPS Global Positioning System
GTS Guaranteed to Send
GUI Graphical User Interface
HR-DSSS High Rate-Direct Sequence Spread
Spectrum
HTN Hybrid Technology Networking
I2C Inter-Integrated Circuit
I2S Integrated Inter-chip Sound
ICMP Internet Control Message Protocol
IDE Integrated Development Environment
IEEE Institute of Electrical and Electronics
Engineers
IP Internet Protocol
IPv4 Internet Protocol version 4
ITU-T International Telecommunication Union-
xviii
Telecommunication
Kbps Kilobits per second
LCD Liquid Crystal Display
LLC Logical Link Layer
LQI Link Quality Indicator
LR-WPAN Low Rate-Wireless Personal Area
Network
MAC Medium Access Control
Mbps Megabits per second
MCPS MAC Common Port Sub-layer
MCU Micro Controller Unit
MHz Mega hertz
MII Media Independent Interface
MIMO Multiple Input Multiple Output
MIPS Million Instructions Per Second
MLME MAC Layer Management Entity
MRDY Master Ready
MSDU MAC Service Data Unit
NLDE Network Layer Data Entity
NLME Network Layer Management Entity
NWK Network
OBMCS On-Board Monitoring and Control
System
OFDM Orthogonal Frequency Division
xix
Multiplexing
PAN Personal Area Network
PCB Printed Circuit Board
PHR Protocol Header
PHY Physical Layer
PID Product Identification
PPDU PHY Protocol Data Unit
RAM Random Access Memory
RFD Reduced Function Device
RFID Radio Frequency Identification
RMII Reduced Media Independent Interface
RSSI Received Signal Strength Indicator
RTS Request to Send
SAP Service Access Point
SAR Successive Approximation Register
SDHC Secured Digital Host Controller
SHR Synchronization Header
SMA Sub Miniature version A
SPI Serial Peripheral Interface
SRDY Slave Ready
SREQ Synchronous Request
SS Station Services
TCP Transmission Control Protocol
TFT Thin Film Type
xx
UART Universal Asynchronous Receiver
Transmitter
UDP User Datagram Protocol
USB Universal Serial Bus
VID Vendor Identification
WEP Wired Equivalent Privacy
Wi-Fi Wireless Fidelity
WLAN Wireless Local Area Network
WSN Wireless Sensor Network
ZCL ZigBee Cluster Library
ZDO ZigBee Device Object
ZNP ZigBee Network Processor
1
Chapter 1. INTRODUCTION TO STATUS
MONITORING IN RAILROADS
1.1. Importance of Freight Railroad
Freight railroad is very important to the economic health of a country as it
moves commodities between source and the market place, between the place of origin
and the place where those commodities are used as raw materials. Hence freight
railroads directly help businesses and industries thrive which in turn drives the
economy of a country.
In the United States of America freight railroad accounts for 40% of the total
freight volume [1], when measured in ton-miles. Ton-mile is defined as the movement
of one ton of freight over a distance of one mile. The commodities moved by freight
rail vary from coal, chemicals, food and related products, automobiles and their
related products, lumber and wood products, minerals, metallic ores, petroleum and
coal and other miscellaneous products [2-3]. Figure 1 and 2 illustrate this distribution.
Figure 1.1. Percent in ton-miles of freight carried by different modes of transport
20%
12%
39% 0%
29%
Percent ton-miles
Pipeline
Water
Rail
Air
Truck
2
Figure 1.2. 2005 Class I Railroads Revenues by Commodity Group
There are several advantages of using freight railroad as well. They are
presented as under:
a) A major advantage of freight rail is that it is fuel efficient. Studies have found
that freight rail is as much as three times as fuel efficient as trucks, the other
major mode of freight movement in the United States.
b) The movement of freight by rail is environment friendly. In today’s day when
the adverse effects of pollution on the environment is all too clear, emphasis is
on clean modes of transport. Environment Protection Agency (EPA) estimates
that freight rail produces only 9 percent of total transportation related nitrous
emissions and 4 percent of total transportation related particulate emissions
although it accounts for nearly 40 percent of inter-city ton-miles [2].
20%
15%
12% 8%
8%
7%
21%
5% 4%
2005 Class I Railroad Revenues by Commodity Group
Coal
Misc. mixed shipments
Chemicals
Motor vehicles & equipment
Farm products
Food & kindred products
All other commodities
Lumber & wood products
Pulp, paper & allied products
3
c) Freight railroad also helps take trucks off the road. It is estimated that a single
freight rail takes 280 – 550 trucks off the road [1]. This contributes
significantly to the alleviation of congestion from the road network.
d) Freight railroad is the safest way to carry hazardous materials. This stems
from the fact that even though both trucks and rail carry hazardous material,
the chances of a truck getting into an accident and releasing hazardous
material is far higher than when using freight rail.
1.2. Reasons for Monitoring Freight Railroad
Each freight railcar consists of hundreds of moving mechanical parts,
electrical wirings, goods and commodities that are prone to damage and maybe
perishable.
The bearing in the wheels of a railcar may overheat and lead to a catastrophic
failure and derailment. Electrical systems may fail leading to shut down of a
refrigeration unit that would spoil the perishable goods that railcar was carrying.
Doors of the railcar may be opened with the intention of stealing merchandise.
Any of these result in huge losses in the form of time and money for both the
railroads and the businesses which depend on the goods transported by the railroads.
Each freight rail consists of close to hundred or more railcars. It becomes very
difficult for the limited number of personnel on-board to monitor all the different
parameter for each and every railcar.
Hence the ability to monitor status of various components and goods in a
railcar becomes of critical importance. Also such monitoring can be used to look at
and evaluate trends that can be used to upgrade maintenance schedules and put best
practices in place.
4
1.3. Existing Methods for Railcar Status Monitoring
Currently employed methods of monitoring railcar status can be classified as
the following:
a) Wayside Monitoring: In wayside monitoring certain parameters of a railcar
are logged when it passes by the detector. The detector is generally placed
outside stations or yards. They are generally used to monitor bearing
temperature which is an indicator of bearing health. The data from the
detectors are transmitted via backhaul networks to controlling stations where
these are monitored and logged. Figure 3 shows a wayside monitoring
installation.
Figure 1.3: Wayside monitoring equipment in the field [5]
b) Automatic Equipment Identification (AEI): This form of monitoring is used
just to identify the railcar or rail equipment that passed by the detector. It
makes use of passive tags that are mounted on the side of railcars or on the rail
5
equipment that is read by a reader mounted by the trackside. This technology
makes use of Radio Frequency Identification technology [6] (RFID). Figure 4
shows an AEI setup in the field.
Figure 1.4: A railcar with an AEI tag mounted on its side [7]
c) On-Board Monitoring and Control System (OBMCS): This system
puts the sensing equipment on-board the railcar. It has two distinct
components. The sensors on-board a single railcar are interconnected by a
CAN bus. While intra-train communication and control from the locomotive is
achieved using IEEE 802.11(b) [8]. The system is also equipped with cellular
radio and GPS for transmission of information to remote monitoring stations
and location services. The system details are present in [9].
6
1.4. Drawbacks of Current Methods
The methods of railcar status monitoring mentioned briefly above suffer from
some severe drawbacks. They are shortly discussed below:
a) Lack of real-time monitoring is a severe drawback plaguing the wayside
monitoring technique. The wayside installations are spaced far apart from one
another. The distance between two stations or yards may be in the hundreds of
miles. Bearing failures are often catastrophic and happen very quickly. There
is a good chance that it may take place between two stations. Also failure of
the refrigeration unit, opening and closing of doors are events that cannot be
effectively monitored by wayside installations.
b) The meager number of parameters that can be monitored by a wayside
installation is also a noticeable drawback. These installations can measure
bearing temperature and some very limited other parameters when a railcar
passes by.
c) The OBMCS addresses these problems but introduces some unique to it. The
use of a wired CAN bus within a railcar fixes the location of the sensors and
also makes it hard to introduce a new sensor later on. The use of cellular radio
is possible only in areas covered by cellular network. There are vast stretches
of land, through which trains need to pass, where there is no cellular coverage.
In those areas this system becomes non real-time. Also, the use of IEEE
802.11(b) [8], cellular radio and GPS [10] make the system power hungry.
The use of dynamo to generate power to recharge the batteries needs invasive
installation on the wheels.
7
1.5. WSN for Railcar Status Monitoring
Wireless Sensor Networks (WSN) involves sensor nodes that are capable of
communicating with other sensor nodes and forming a network among themselves
[11]. These nodes are capable of monitoring parameters and then communicating the
readings through the network to the main collecting unit. Since the nodes are
extremely resource constrained so the design of sensor nodes and network protocols
are all guided by the principle of low power consumption. The continuing
advancement in silicon science enables reduction of size of the sensor nodes which
leads to lowering of cost. WSN has been successfully used in varied application
environments as follows:
a) Area Monitoring
b) Environmental monitoring of forests, glaciers, air quality etc.
c) Industrial monitoring of machine health, data logging etc.
d) Agricultural monitoring of crops, soil humidity, cattle etc.
e) Smart home monitoring.
The deployment of WSN for railcar status monitoring is hence very feasible. It
allows for real-time or near real-time monitoring. The absence of wires and self-
forming and self-healing networks make the addition of new sensors very easy. Low
power consumption profile increases the overall lifetime of the network and also since
these can easily be battery powered, invasive installation is not required.
Hence the use of WSN for railcar status monitoring has significant advantages
over the currently used methods of monitoring.
8
Chapter 2. OVERVIEW OF COMMUNICATION
PROTOCOLS
This chapter will provide a brief overview of the involved communication
protocols, namely IEEE 802.11, IEEE 802.15.4 and ZigBee.
2.1. IEEE 802.11
IEEE 802.11 [12] consists of a group of standards that are used to implement
Wireless Local Area Network (WLAN) communication in the 2.4, 3.6, 5.0 and 60.0
GHz frequency bands. In 1999, the Wi-Fi Alliance [13] was formed which is a body
that evaluates and certifies a product as Wi-Fi compliant. Wi-Fi is the name that has
come to be commonly associated with products that conform to the IEEE 802.11
group of standards.
The following table shows the evolution of the IEEE 802.11 standard over the
years with respect to a few parameters of interest:
802.11
Protocol
Release
Year
Frequency
(GHz)
Bandwidth
(MHz)
Data Rate
(Mbps)
Modulation
Initial 1997 2.4 20 1,2 DSSS,
FHSS
a 1999 5.0 20 6,9,12,18,24
36,48,54
OFDM
b 1999 2.4 20 1,2,5.5,11 DSSS
g 2003 2.4 20 6,9,12,18,24
,36,48,54
DSSS,
OFDM
9
n 2009 2.4, 5 20 7.2, 14.4,
21.7, 28.9,
43.3, 57.8,
65, 72.2
OFDM
40 15, 30, 45,
60, 90, 120,
135, 150
Table 2.1: IEEE 802.11 releases over the years [8]
2.1.1. IEEE 802.11 Components
The IEEE 802.11 typical network architecture consists of the following
components:
a) Basic Service Set: The IEEE 802.11 network is envisioned to consist of
several interconnected cells. Each cell is called the Basic Service Set.
b) Access Point: Each Basic Service Set is to contain an Access Point or Base
Station which controls the network within that cell.
c) Distribution System: If the WLAN consists of several cells, then the APs are
connected by either a wired or wireless backbone. This is known as the
distribution system.
d) Extended Service Set: The entire network consisting of the Basic Service
Sets, APs and distribution system, viewed from the higher layers of the OSI-
ISO stack appear as a single entity known as the Extended Service Set.
10
e) Portal: This is defined as a device which connects the 802.11 LAN to any
other 802 LAN.
Figure 5 shows the conceptual diagram of an 802.11 WLAN consisting of
most of the above components.
Figure 2.1: The IEEE 802.11 Infrastructure based network architecture [14]
2.1.2. IEEE 802.11 Network Types
The IEEE 802.11 standards provide for two distinct of network modes. The
two modes are discussed as follows:
a) Infrastructure mode: This type of network depends on the cellular
architecture and the presence of an Access Point through which all subscriber
stations within a cell connect to the network. Figure 5 is an ideal
representation of this type of network. The Distribution System or DS in an
infrastructure based network needs to provide Distribution system services
11
(DSS) and Station services (SS). Distribution system services are related with
node mobility and include Association, Re-association, Disassociation,
Distribution and Integration. The Station services include Authentication, De-
authentication, Privacy and MAC Service Data Unit (MSDU) delivery.
b) Ad-hoc mode: In this mode there is no Access Point present. There are just
nodes which form a network among themselves. The services that are present
in the Infrastructure mode are taken up by end-user stations, like beacon
generation. Some functions, like frame relaying and power saving, available in
the Infrastructure mode are not available in the ad-hoc mode. Figure 6 shows
the layout of a typical ad-hoc network.
Figure 2.2: A few laptops in an ad-hoc IEEE 802.11 network [15]
12
2.1.3. IEEE 802.11 Protocol Stack
Figure 7 shows the IEEE 802.11 protocol stack showing the components of
the Physical and the Data Link Layers.
Figure 2.3: IEEE 802.11 Protocol Stack [16]
2.1.3.1. IEEE 802.11 Physical Layer
The original IEEE 802.11 standard defined communication at 1 or 2 Mbps
using Frequency Hopped Spread Spectrum (FHSS) or Direct Sequence Spread
Spectrum (DSSS). Later on Orthogonal Frequency Division Multiplexing (OFDM)
and High Rate-Direct Sequence Spread Spectrum (HR-DSSS) were introduced to
support higher data rates.
When the FHSS technique is used the 2.4 GHz band is divided into 75 1 MHz
sub-channels. Each pair of sender and receiver agrees on a hopping pattern and
conversation between them is carried out with communication hopping between sub-
channels according to the agreed hopping pattern. Each pair also chooses the hopping
pattern in a way that interference is minimized. FHSS technique has the following
characteristics:
13
a) Radio design is simple.
b) Data rate limited to 2 Mbps due to sub-channel restriction to 1 MHz by FCC.
c) Frequent hopping leads to high hopping overhead.
The use of DSSS technique allows the use of 13 22 MHz sub-channels in the
2.4 GHz band. All of these bands are partially overlapping with one another. Hence at
any time there are just 3 sub-channels which are completely overlap free and
communication is carried over any one of these channels. DSSS technique has the
following characteristics:
a) Data rate is no longer limited to 2 Mbps.
b) Hopping is not used.
c) Chipping is used to increase immunity to noise. Each bit in the transmission
sequence is converted into a unique series of bits called a chip and the entire
signal is spread over the 22 MHz band to reduce the effect of noise and also
build is redundancy that can be used for error correction. This leads to fewer
retransmissions.
Figure 8 shows the utilization of the 2.4 GHz band using DSSS technique.
Figure 2.4: Channel assignment in the 2.4 GHz band for IEEE 802.11 [8]
14
2.1.3.2. IEEE 802.11 Data Link Layer
The IEEE 802.11 Data Link Layer consists of the Logical Link Control (LLC)
sub-layer and the Medium Access Control (MAC) sub-layer.
The LLC is the same for IEEE 802.11 as it is for any other 802
communication protocol [17].
The MAC is unique to the IEEE 802.11. It provides the following features:
a) Distributed Coordination Function (DCF) deals with the problem of
collision when two or more stations try to simultaneously transmit over the
same channel. Carrier Sense Multiple Access/Collision Avoidance
(CSMA/CA) is used for this purpose. When a station has data to transmit, it
senses the channel and waits for a randomly generated time interval, if it finds
the channel free. At the end of this time period if the channel is still free then
the transmission is made. The receiving station receives the packet and if the
CRC is alright, sends an acknowledgement back to the sender. If the sender
receives the acknowledgement transmission terminates successfully. If the
packet was lost or received in error or the acknowledgement was lost, the
sender waits for some time and sends the packet again for a specified number
of retries.
b) Hidden node problem: Often times there may arise a situation in which both
the sending and receiving stations are visible to the Access Point but are
outside of range from each other. In these cases the sender sends a Request to
Send (RTS) to the access point. Since the Access Point is visible to all nodes
in the cell, a Clear To Send (CTS) is only sent to the sender when there is no
15
chance of collision happening during the ensuing transmission. This way the
hidden terminal problem is tackled.
c) Fragmentation and Reassembly: Large packets are always more susceptible
to corruption by noise and other inherent disturbances in the wireless channel.
The IEEE 802.11 MAC layer provides for breaking a large packet into smaller
chunks each framed by a MAC header and CRC for better reliability in
transmission. On the receiving end it is the responsibility of the MAC layer for
reassembling the fragments before passing them onto a higher layer.
2.1.4. IEEE 802.11 Security
Transmission security is one of the foremost concerns in any deployment of
the IEEE 802.11 networks. Since it is over-the-air the ease of eavesdropping is much
easier than in wired networks. The IEEE 802.11 standard defines a Wired Equivalent
Privacy (WEP) for the security of communication. This has been shown to be very
weak in the face of sophisticated attacks [18-20].
In summary, IEEE 802.11 provides for a high data rate long range
communication protocol. The finer details of the protocol and its various
implementation details can be found in [12].
2.2. IEEE 802.15.4
IEEE 802.15.4 [21] standard specifies the physical layer and the medium
access layer for Low Rate-Wireless Personal Area Networks (LR-WPAN). It is
designed for networks where communication is within devices close by and
infrastructure for network formation and maintenance is virtually absent. It focusses
on low power consumption, low data rate and small area of coverage. It is in sharp
contrast to IEEE 802.11 or Wi-Fi where the networks deployed are principally
16
infrastructure based and power consumption is not an issue but end-user experience is
[22].
IEEE 802.15.4 was initially theorized to have a communication distance of 10
m and a maximum theoretical data rate of 250 kbps.
The two versions of IEEE 802.15.4 that are of interest in this research work
and their brief introduction is provided in the table below:
IEEE 802.15.4 version Brief Introduction of features
IEEE 802.15.4 – 2003 This was the first release of the IEEE
802.15.4 standard. It provided two
physical layer bands, one in the 868 and
915 MHz and the other at 2.4 GHz.
IEEE 802.15.4 – 2006 This version of the standard upgraded
the data rates possible in the 868 and
915 MHz bands. It also introduced new
modulation schemes - three for the 868
and 915 MHz and one for the 2.4 GHz
band.
Table 2.2: IEEE 802.15.4 versions [23]
2.2.1. LR-WPAN Device Architecture
The LR-WPAN device architecture is shown in figure 9 below. The IEEE
802.15.4 standard specifies the physical and the MAC layers only. The upper layers
have to be implemented according to usage scenarios.
17
Figure 2.5: LR-WPAN Device Architecture [24]
The details of the layers and their interactions are provided in [21,25].
There are some popular upper layer implementations using IEEE 802.15.4 as
the base and they include, but are not limited to, ZigBee [26], 6LowPAN [27], ISA-
100.11a [28] and MiWi [29].
2.2.2. IEEE 802.15.4 Physical Layer
The physical layer of IEEE 802.15.4 specifies the radio communication
capabilities of the protocol. The standard specifies several frequency ranges of
operation, data rates and modulation techniques which are tabulated in the following
table:
18
Frequency
(MHz)
Available
Channels
Bit Rate
(kbps)
Symbol
Rate
(kbaud)
Modulation
Scheme
Receiver
Sensitivity
(dBm)
868 – 868.6 1 20 20 BPSK -92
902 - 928 10 40 40 BPSK -92
868 – 868.6 1 250 12.5 ASK -92
902 - 928 10 250 50 ASK -92
868 – 868.6 1 100 25 O-QPSK -92
902 - 928 10 250 62.5 O-QPSK -92
2400 –
2483.5
16 250 62.5 O-QPSK -86
Table 2.3: IEEE 802.15.4 Frequency bands of operation [30]
The physical layer of IEEE 802.15.4 provides two distinct services:
a) Data transmission and reception using physical layer protocol data units
(PPDU).
b) Interface to the physical layer management entity which provides access to the
upper layers as well as maintains information about the WPAN.
Direct Sequence Spread Spectrum technique is used for transmission which
enhances the robustness of the communication in a non-perfect channel.
The data exchange is carried through using PPDUs the format of which is
presented in the figure below:
19
Figure 2.6: Format of a IEEE 802.15.4 PPDU [21]
The SHR is used by the receiving device for synchronization, the PHR
contains the frame length information and the PHY payload is variable length and
contains the packet that is handed down from the IEEE 802.15.4 MAC layer.
The management functions that the IEEE 802.15.4 physical layer carries out
have been listed below:
a) Receiver Energy Detection (ED): This is used by the network layer as a part
of its channel selection algorithm.
b) Link Quality Indication (LQI): This is a representation of the strength and
quality of the received packet. This is passed on to the higher layers and its use
is completely dependent on the higher layers.
c) Clear Channel Assessment: Clear channel assessment is done using any one
of the three methods outlined hereafter. If the channel sense returns energy
above ED threshold then the channel is deemed to be busy. If on sensing the
channel a carrier with IEEE 802.15.4 physical layer characteristics and
spreading profile is detected then the channel is deemed to be busy. The last
method is to use a combination of carrier sense and also detecting whether that
energy is above ED threshold, to decide if the channel is busy or not.
20
2.2.3. IEEE 802.15.4 MAC layer
The IEEE 802.15.4 MAC layer provides the interface between the physical
layer and the upper layers. Since the upper layers are not defined as part of the
standard, they can be any of the implementations, like ZigBee, 6LoWPAN, mentioned
before. The IEEE 802.15.4 MAC layer provides the following services [21]:
a) MAC Data Services: The entity known as MAC Common Port Layer
(MCPS) provides data transmission and reception services between peer MAC
layers.
b) MAC Management Services: The MAC Layer Management Entity (MLME)
provides the interface through which layer management functions are
accessed. It also maintains a database of objects for the MAC layer known as
the Personal Area Network (PAN) information database. The MLME also has
access to the MCPS for data transfer activities.
In general the IEEE 802.15.4 MAC layer is involved in the following network
functions:
a) Beacon management.
b) Channel access.
c) Guaranteed-to-Send (GTS) management.
d) Frame validation.
e) Acknowledgement frame delivery.
f) Association
g) Dissociation.
The general MAC layer frame format is presented below and a short
discussion is given thereon about the various fields within the frame.
21
Figure 2.7: IEEE 802.15.4 MAC Frame format [21]
The MAC frame consists of the MAC header (MHR), a variable MAC
payload and a Frame Check Sequence (FCS).
The MAC header consists of information needed for proper traversal of the
data payload through the physical network to the intended receiver device. It consists
of the following fields:
a) Frame Control: A 2 bytes long frame control field contains information
related to frame type, addressing fields and other control flags. This will be
discussed in details subsequently.
b) Sequence number: A 1 byte long sequence number gives an identifier to the
frame that will be used by the destination node to generate an
acknowledgement frame with the same sequence number.
c) Destination PAN identifier: A 2 bytes long field, when present, indicates a
unique PAN identifier of the destination node.
d) Destination address: A 2 or 8 bytes long field, when present, that specifies
the address of the intended recipient of the frame.
e) Source PAN identifier: A 2 bytes long field, when present, that specifies the
unique PAN identifier of the origin of the frame.
22
f) Source address: A 2 or 8 bytes long field, when present, that specifies the
address of the origin of the frame.
g) Auxiliary Security Header: This field provides all the information necessary
for security processing of the frame.
h) Frame payload: This field contains the data to be transmitted as handed down
by the upper layers at the transmitter. If security is enabled then this field may
be cryptographically protected.
i) FCS field: This field is a 16-bit ITU-CRC that is calculated over the entire
MHR and MAC payload fields.
The structure of the Frame control field is shown in the figure below and its
various bits and their functions are discussed.
Figure 2.8: IEEE 802.l5.4 MAC Frame control field [21]
The Frame control field bits are used as follows:
a) Frame type: Bits 0 and 1 are used to indicate the frame type. There are four
types of frames namely Beacon, Data, Acknowledgement and MAC
command.
b) Security enabled: This field if set to 1 indicates that the data in the payload
will be encrypted and the auxiliary security header will be present in the MAC
frame.
23
c) Frame pending: This bit will be set to 1 if the transmitter of the frame has
more frames to send after the current one.
d) Acknowledgment request: This bit when set to 1 indicates that the recipient
of the data or MAC command frame needs to acknowledge the receipt. An
acknowledgement frame with the appropriate sequence number is sent back in
that case.
e) PAN ID compression field: This bit indicates whether only one of the PAN
identifiers will be present in the frame even though both source and
destination addresses are present. If this field is 1 then when both source and
destination addresses are present, only the destination PAN identifier will be
present and the source PAN identifier will be assumed to be equal to the
destination PAN identifier. If this field is set to 0, then the PAN identifier will
be present only if the corresponding address fields are present.
f) Destination and Source addressing modes: These fields can be set to any
one of the following values:
PAN identifier and address fields are not present (00)
Address field contains a short address (10)
Address field contains an extended address (11)
Reserved (01)
g) Frame version: The presence of a 0 in this field indicates that the frame is
compatible with IEEE 802.15.4-2003 and a value of 1 indicates just an IEEE
802.15.4 frame. All other values are reserved.
The transfer of data between devices using IEEE 802.15.4 takes place in one
of two modes. These are known as the Beacon Enabled mode and the non-Beacon
24
Enabled mode. The beacon enabled mode uses the slotted CSMA/CA algorithm
whereas the non-Beacon enabled mode uses an un-slotted CSMA/CA algorithm [21].
The sequence of data transfer operations using these two modes are presented in the
figures below.
Figure 2.9: Sequence of steps when Coordinator has data to send to Network
device in a beacon enabled mode [21]
Figure 2.10: Sequence of steps when the Network device has data for
Coordinator in a beacon enabled mode [21]
25
Figure 2.11: Sequence of steps for transfer of data in a non-beacon enabled
network [21]
2.2.4. IEEE 802.15.4 Network Topologies
There are two types of network topologies that can be used within the IEEE
802.15.4 standard. These are briefly defined below [31]:
a) Star Network: In this topology there is a central PAN coordinator node and
every other node communicates with this node.
b) Peer-to-Peer Network: In this topology there still exists the PAN coordinator
node but the other nodes can also communicate between themselves without
having to go through the coordinator node.
The two network topologies are represented in the figure below:
26
Figure 2.12: IEEE 802.15.4 network topologies a) Star and b) Peer-to-Peer
2.2.5. IEEE 802.15.4 Device Types
The different types of device classes that can exist within an IEEE 802.15.4
network are briefly presented below:
a) Full Function Device (FFD): These devices can send, receive as well as route
data through them. They require more power to function.
b) Reduced Function Devices (RFD): These are devices which can only send
and receive data but cannot route traffic through them. They are generally end-
devices in a network, typically a sensor or a switch. They can be very low
power devices as they do not need to route traffic and hence can be out to
sleep when not in operation.
27
c) PAN Coordinator: This is a special type of FFD which creates the network
and also coordinates and manages it.
2.3. ZigBee
ZigBee is a standard which defines the network and application layer over the
IEEE 802.15.4 physical and MAC layers for wireless sensor networks. It is
maintained by ZigBee alliance, which an open, non-profit collaboration of hundreds
of companies. Before any device is declared ZigBee compliant it has to undergo the
ZigBee Certification Program [32]. ZigBee has the following characteristics:
a) Leads to low power consumption hence longer battery life.
b) Leads to small device footprint.
c) Allows mesh networking.
d) Self-forming, self-healing network.
e) Interoperability between devices from various vendors but conforming to the
ZigBee standard.
f) Security mechanisms in place with 128 bit AES encryption possible.
In addition to the standard the ZigBee Alliance has defined many application
profiles as well. Every data request in ZigBee is sent and received on an application
profile. Some of the commonly used application profiles are Home Automation,
Industrial Plant Monitoring, Commercial Building Automation, Telecom applications,
Personal Home and Hospital care etc [33].
2.3.1. ZigBee Stack Architecture
The best way to understand how ZigBee builds on the IEEE 802.15.4 lower
layers is by looking at the ZigBee stack architecture as provided in [33]. The same is
replicated in the figure below:
28
Figure 2.13: ZigBee stack architecture
Each component of the ZigBee standard will be briefly discussed in what
follows
2.3.1.1. ZigBee Network Layer (NWK)
The ZigBee network layer resides just above the MAC layer as defined by
IEEE 802.15.4 and is in charge of executing the following services [34]:
a) Routing of frames.
b) Discovering and maintaining route tables, one-hop neighbors and storing
important information about neighbor devices.
c) Joining and Leaving of ZigBee network devices.
d) Providing cryptographic security to the transmitted frames.
29
The above functions are achieved by the NWK layer using the NWK Layer
Data Entity (NLDE), used for data transmission via the NLDE-Service Access Point
(NLDE-SAP), and the NWK Layer Management Entity (NLME) that manages the
services via the NLME-Service Access Point (NLME-SAP).
The general NWK layer frame format is presented below and each of the
fields in the frame are defined.
Figure 2.14: General ZigBee NWK Layer Frame Format [34]
The fields in the NWK Layer Frame are presented briefly below:
a) Frame Control: This field has information about how the frame is going to be
treated. Each bit within the Frame control field has a specific meaning as
follows:
Frame type: This indicates the type of frame like NWK data or NWK
command etc.
Protocol Version: The value here indicates which version of ZigBee
this frame is compliant with. For example, a value of 2 will indicate
ZigBee 2006/2007/Pro.
30
Discover Route: If this contains the value 1 then route discovery for
this frame is enabled.
Multicast Flag: If this bit is 0 then it indicates that the frame will be
unicast to its destination.
Security: If this bit is 0 then security for the frame is disabled and if it
is 1 then security for the frame is enabled.
Source Route: If this field is 0 then it means that the source route is not
specified for the frame.
Destination and Source IEEE addresses: A zero in each of these fields
means that the frame does not contain the field marked with a zero. A
1 will indicate that the corresponding field is present in the frame.
b) Destination Address: This field contains the 16 bits long destination short
address.
c) Source Address: This field contains the 16 bits long source short address.
d) Radius: This field contains the range of transmission as the maximum number
of hops from source to destination. Each node on the way decrements this
value by 1.
e) Sequence number: This is a number unique to the frame being sent and is
used for proper re-assembly at the receiver end.
f) Source and Destination IEEE Address: If present these fields contain the 64
bits long source and destination IEEE addresses respectively.
g) Multicast Control: This field is present dependent on the Multicast flag in the
Frame Control field.
31
h) Source Route sub-frame: This is present if the Source route flag in the Frame
control field is set to 1. This indicates to the next hop node to check this sub-
frame to determine the next hop in the route to the destination for this frame.
i) Frame payload: This contains the data that has been handed down from the
higher layers.
2.3.1.2. ZigBee Application Layer
The application layer of ZigBee is composed of Application Framework (AF),
ZigBee Device Object (ZDO) and Application Support Sub-layer (APS) [34].
The Application Support Sub-layer (APS) lies between the application and
network layers and principally accomplish the following tasks:
a) It uses the APS Data Entity (APSDE) for exchange of data.
b) It uses the APS Management Entity (APSME) for exchange of management
primitives.
c) The APS Information Base (AIB) contains constant and variable attributes.
d) When the APS receives an application payload, it adds the address and control
fields, and passes the resulting payload to the lower layer.
e) APS frames are of the following types namely data, command, and
acknowledge.
The Application Framework (AF) has the following characteristics:
a) Each device can have up to 240 application endpoints represented by a 8 bit
address.
32
b) The application endpoint is used to transmit and receive data. It acts much like
a port in TCP/IP communications.
c) Information generated by an application endpoint is called an attribute. A
collection of attributes and commands working on the attributes is called a
cluster. A collection of clusters is called the ZigBee Cluster Library (ZCL). A
profile consists of a cluster, device ID and endpoint and together it forms an
application. Each attribute, cluster and profile is represented by 16 bit IDs.
The ZigBee Device Object (ZDO) layer has the following functions to
perform:
a) Service and device discovery.
b) Initialization of the coordinator.
c) Security management.
d) Binding management.
e) Network management.
The frame formats involved in the discussed layers and details about the layer
functionalities can be found in [33].
2.3.2. ZigBee Device Types
ZigBee standard defines the following device types:
a) Coordinator: It is responsible for starting and maintaining the network. It is
also responsible for assigning addresses to newly joined devices. It is a FFD as
defined in the IEEE 802.15.4 standard.
b) Routers: These devices are used to route frames in the network and are likely
to be always powered on. These are also FFDs as defined in the IEEE
802.15.4 standard.
33
c) End-Devices: These are the same as RFDs as defined in the IEEE 802.15.4
standard.
2.3.3. ZigBee Network Topologies
ZigBee standard defines the following network topologies for use:
a) Star topology: In this topology the central PAN coordinator node starts and
maintains the network. Every other device can only communicate with and
through the PAN coordinator.
b) Tree topology: In this topology beacon enabled mode is used in a hierarchical
communication structure.
c) Mesh topology: In this topology peer-to-peer communication is allowed. This
leads to a network that is more robust to failures because of the existence of
redundant paths.
The Star and Mesh topologies are very much similar to the Star and Peer-to-
Peer topology shown during the discussion of IEEE 802.15.4 topologies. The Tree
topology introduced in ZigBee is shown in the figure below.
Figure 2.15: ZigBee Tree Topology [35]
34
Chapter 3. ISSUES WITH ZIGBEE IN RAILROAD
ENVIRONMENTS
The North American Railroads are looking at the use of on-board real-time or
near real-time monitoring of freight railcar status using WSN. ZigBee is the
technology of choice for the railroads for its low power consumption and low cost of
implementation.
In the railroad environment typically there will be at least hundred railcars
strung together to form a freight rail. Each of the railcars will be equipped with
several sensors. Let us assume that each railcar has only one sensor node. Then the
data from each of these sensor nodes will have to travel hop-by-hop to the monitoring
station in the locomotive. So the network topology that is most commonly found in
the railroad environment is a long linear-chain like topology [36].
ZigBee was primarily designed for star topology [37]. Tests carried out at the
Advanced Telecommunication Laboratory at the University of Nebraska, has shown
that although ZigBee supports multi-hop communication but it is unsuitable for
topologies involving as many hops as in the railroads environment [38 – 40].
Similar research studies carried out by others have also found similar issues
and performance degradations with ZigBee [41] and also link layer performance
bottlenecks as found in [42].
In the following pages we briefly revisit the main issues and challenges that
were found to plague a ZigBee-only network in the railroads environment.
35
3.1. Route Discovery
A route in a network, wired or wireless, refers to the path that packets take
from their origin to the intended destination. Each link from one node to its next
communication partner is referred to as a hop. If a router has more than one hop, we
refer to this as multi-hop routing [36].
ZigBee was designed for star topologies. In such a configuration, each node is
in direct communication range of the PAN Coordinator. In this topology the hop
count is at most 2 hops. To a limited degree, ZigBee also supports tree topologies,
with the PAN Coordinator as the root node. However, it is severely restricted in the
number of hops it supports for this topology. A hop count of 100 or more is far
beyond the capability of ZigBee. In the chain-like topology we find on a freight train,
ZigBee fails to discover routes after about 20 – 25 hops [36].
Figure 3.1: Number of unsuccessful route requests vs Node ID
3.2. Synchronization Delay
All nodes within a single PAN network are synchronized with each other.
Each time the layout of the network changes due to node failure or disconnections, a
resynchronization becomes necessary. This can also be triggered by fluctuating
0
20
40
60
80
100
120
140
160
0 10 20 30 40 50
Ro
ute
Re
qu
est
s
Node ID
Route Requests
36
channel conditions that prevent a node from successfully communicating with the rest
of the network for a period of time. Fluctuating channel conditions are a common
occurrence in wireless networks. For a 100 railcar scenario, this synchronization can
stretch to several minutes. During this time, no communication is possible between
the unsynchronized nodes and the rest of the network [36].
Figure 3.2: Time required for synchronization vs the maximum number of
hops in network
3.3. Packet and Link Loss
If we study the behavior of wireless communication systems we can find that
the probability of successful packet delivery decreases as the number of hops
increases. As reported in [41], our analysis of this problem when applied to freight
train monitoring using ZigBee. That analysis is shown in the following figure,
demonstrating the end-to-end packet error rate for different hop counts and bit error
rates. For this analysis we used a packet size of 127 bytes per packet [36].
0
50
100
150
200
250
300
1 11 21 31 41 51 61 71 81 91Syn
chro
niz
atio
n D
ela
y (s
eco
nd
s)
Number of Hops
Synchronization Delay
37
Figure 3.3: Packet Loss Rate percentage for different hop counts and Bit
Error Rates
It can be seen, for a Bit Error Rate of 10-5
, which in wireless networks is a
very low BER signifying good channel conditions, we still experience a packet loss of
over 60% due to errors. Less than 40% of packets are being delivered successfully for
a hop count of 100. Under more realistic conditions, with BERs between 10-3
and 10-
4, we can observe that none of the packets from a majority of nodes on the train will
successfully be delivered to the locomotive. This means that the network will
completely cease to function.
3.4. Lack of Quality-of-Service
Quality-of-Service refers to the ability of a network to allow transport of
packets with special requirements, and appropriately allocate transmission resources
to it. ZigBee does not have a way of assigning priorities to different packets.
This is extremely important in the freight train monitoring system. An
important aspect of the envisioned system is the ability to generate alerts in case of
equipment failure, the detection of dangerous substances or leaks, etc. But if these
0 10 20 30 40 50 60 70 80 90 1000
10
20
30
40
50
60
70
80
90
100
Number of Hops
Pac
ket
Loss
Rat
e (p
erce
nts)
Packet Loss Rate in Multi-hop Network
BER 10-5
BER 10-4
BER 10-3
38
alert messages are treated the same way as a message containing periodic status
updates from, say, a refrigeration unit, then it shares the same probability of being
discarded by the network due to errors or congestion. It also will be entered into every
queue at the end, like all other packets, and thus be significantly delayed. For alerting,
this is simply not feasible [36].
3.5. Data Forwarding and Aggregation
A chain topology using ZigBee means that data from the last node needs to be
relayed to the first node. In addition to relaying that information, however, each node
also has its own data to report. Hence the amount of data quickly aggregates. By the
time the nodes closest to the destination are reached the volume of data becomes very
large, essentially overwhelming the network and causing significant congestion. This
leads to excessively high packet loss. Simulation shows that if a freight train consists
of more than 25 railcars then it cannot even sustain traffic of small 50-byte message
per node generated every 20 seconds [36].
Figure 3.4: Amount of data present in channel as a progression of hops from
the end of train
Overflow
(lost data)
Channel
Throughput
New Sensor Data Forwarded Data Lost Data Received Data
39
3.6. Network Lifetime
The nodes closest to the data destination have a lot more data to forward than
nodes towards the end of the train. Hence, their increased number of operations also
means that they consume a lot more energy and deplete their reserves much faster.
Due to the point-of-failure issue, if one of those nodes fails, no data from behind that
node will be able to reach the destination anymore. At this point, the network
effectively fails to operate [36].
Figure 3.5: Energy consumption per node for a 100-railcar train
The battery life profile illustrates the claim that the nodes closer to the
coordinator node run out of battery faster than the ones farther away from it. We
can see that, although node 0 would be operational for around 2000 hours, the
network will fail after less than 200 hours [36].
0 10 20 30 40 50 60 70 80 90 1000
5
10
15
20
25
Multi-hop Sensors Per Node Energy
Node ID
Ene
rgy
cons
umpt
ion
(Jou
les)
40
Figure 3.6: Battery life in hours for each node in a 100-railcar train
0 10 20 30 40 50 60 70 80 90 1000
200
400
600
800
1000
1200
1400
1600
1800
2000
Battery Life in Multi-hop Sensors
Node ID
Tim
e (H
ours
)
41
Chapter 4. OVERVIEW OF HYBRID
TECHNOLOGY NETWORKING (HTN)
The Advanced Telecommunication Engineering Laboratory at the University
of Nebraska-Lincoln has come up with a solution that addresses the issues that a
ZigBee-only network faces in the railroads application environment. The solution is
named Hybrid Technology Networking (HTN) protocol [43].
The basic idea behind the protocol envisions railcar monitoring as a two-step
process.
In the first step clusters of sensors are formed spanning the length of the
freight rail. Each of the clusters have only one coordinator node which is called the
gateway. Each sensor node communicates via single or limited number of hops to the
gateway. In the next step, the gateway aggregates the sensor data and relays it to the
locomotive via Wi-Fi hops over similar gateways spanning the length of the freight
rail [43].
The high-level overview of the Hybrid Technology Networking protocol is
shown in the figure below:
Figure 4.1: High-Level representation of Hybrid Technology Networking
42
Within the clusters ZigBee is used for communication. Since the clusters are
small and of low depth the synchronization delay is drastically reduced.
Figure 4.2: Time required for synchronization vs. the maximum number of hops
in network in HTN
Since the ZigBee networks operate in small clusters the chance of the network
getting overwhelmed due to data aggregation is also minuscule.
Figure 4.3: Per Node Throughput vs. the maximum number of hops in HTN
0
50
100
150
200
250
300
1 11 21 31 41 51 61 71 81 91
Syn
chro
niz
atio
n D
ela
y (s
eco
nd
s)
Number of Hops
Synchronization Delay with HTN
0
500
1000
1500
2000
2500
3000
3500
4000
1 11 21 31 41 51 61 71 81 91
Pe
r-N
od
e T
hro
ugh
pu
t (k
bp
s)
Number of Hops
Per-Node Throughput with HTN
43
As can be seen in Figure 28 the per node throughput using ZigBee-only
network is shown with the red bar, the improvement in throughput with HTN is
clearly visible.
The protocol also suggests rotation of the gateway role among all the capable
nodes in a cluster hence balancing the power consumption profile.
Figure 4.4: Node Lifetime vs. Total Number of Hops in HTN
The above simulation result clearly shows the improvement in network
lifetime due to a balanced power consumption profile.
As mentioned in [43], simulation results with HTN shows that there are the
following significant advantages over a ZigBee-only network in the railroads
application environment:
a) 90% reduction is synchronization delay.
b) 40 times improvement in system throughput.
c) Increasing network lifetime by about 250%
d) Since HTN protocol sits on top of the underlying ZigBee and Wi-Fi protocols,
an end to end Quality-of-Service mechanism can be implemented.
0
500
1000
1500
2000
1 11 21 31 41 51 61 71 81 91
No
de
Lif
eti
me
(h
ou
rs)
Number of Hops
Node Lifetime with HTN
44
Chapter 5. PROBLEM STATEMENT
This chapter underscores the reason behind undertaking this research work.
As is evident from the discussions in chapter 3 the application of ZigBee-only
networks to monitor railcar status is nowhere near an optimal solution. It suffers from
severe drawbacks which defeat the very purpose of using WSN to monitor railcars.
A solution is presented in chapter 4 and it is seen that this solution addresses
all of the issues faced by a ZigBee-only network in the application scenario. The
performance improvements obtained from using this new protocol, called HTN
protocol, is also evident from the simulation results presented by the authors.
The next step in exploring this solution is to implement it in hardware and
actually deploy WSN using the HTN protocol and evaluate the results obtained from
the field tests.
The first step towards this goal is to find a sensor platform that is capable of
supporting HTN. The major components of such a sensor platform will be a Wi-Fi
and a ZigBee radio available on the same platform.
The available sensor platforms like MICAz [44], Stargate [45] and recent
development Libelium Waspmote [46] are all capable of providing the ZigBee
communication part as well as the Wi-Fi communication part but not in an integrated
solution. The current available sensor platforms will have to be plugged into one
another to realize both the communication radios together.
Several technical and logistical problems arise from this approach. Some of
these problems are listed below:
a) The data rate of plug-in boards will always be limited by the bus that is used to
connect them. In all of the above cases, except the Stargate, the only available
45
interfacing bus is the UART. This is slow speed and hence will cause a
bottleneck for Wi-Fi traffic. In Stargate, the PCMCIA interface is used which
removes the speed bottleneck but is an order of magnitude more power hungry
than the UART.
b) The cost of obtaining various pieces of hardware from different vendors,
assembling them and coding for the required firmware is high and is a
logistical nightmare when the numbers of such platforms required are in
millions.
c) The railroads are looking for standards compliant solution so that vendor
independence is maintained. Some of the off-shelf solutions may not be
standards compliant.
d) A custom hardware can be made with specific design goals in mind. The
commercially available components were designed with some design goals
and when they are brought together they may not meet the goals of the
railroads.
As can be understood from the above discussion there are major advantages in
designing a custom integrated hardware platform that will be standards compliant and
also will meet and exceed any design goals that may be set by the railroads.
46
Chapter 6. HARDWARE DESIGN
The purpose of the hardware designed here-in is to provide an integrated
platform for the implementation of the HTN protocol. Chapters 3 has shown the
issues related with the use of ZigBee-only networks in the railroads scenario. Chapter
4 discusses about a solution that addresses the issues highlighted in chapter 3. In
chapter 5 we provide the reasons why a new sensor hardware platform is necessary
for the implementation of the HTN protocol.
In this chapter the hardware design has been presented in details. The
operational requirements for the hardware have been discussed, followed by a general
high level block diagram of the hardware. This is followed up by a discussion on each
of the components involved in the design with respect to why they were chosen, how
they are interfaced with the microcontroller etc. The design choices made during the
design of the PCB have been discussed. Several snapshots of the hardware have been
provided at the end of the section.
6.1. Requirements
The hardware design was guided by a set of requirements which originated
from the overview of the HTN protocol and what it wishes to achieve.
The requirements are listed below:
a) The microcontroller that will form the core of the hardware platform needs to
be fast as it has to deal with both ZigBee and Wi-Fi communication and also
run a sizeable firmware involving many calculations. The microcontroller
should be able to drive data transfer to and from ZigBee radio to Wi-Fi radio
and vice-versa fast enough to ensure respectable throughput. At the same time
47
the peripherals available with the microcontroller should be large. Keeping the
overall low power budget in mind the microcontroller and its associated drive
circuitry should not consume much power.
b) ZigBee radio hardware. The best option is to have an integrated solution
having an on-board processor that will offload the ZigBee communication
related tasks from the microcontroller. Needs to be able to transmit at a max of
3dBm output power and have a receive sensitivity of up to -87 dBm. The
overall power consumption of this component should not exceed the 50 mA in
any communication mode (Tx/Rx). Should operate in the 2.4 GHz band and
have the flexibility to pick and cycle between the 16 available channels.
c) Wi-Fi radio hardware. The best option is to have an integrated solution having
an on-board processor that will offload the TCP/IP stack functionality and
radio interface tasks from the microcontroller. Should be low on power
consumption, not exceeding 150 mA in any of the communication modes
(Tx/Rx). Should be compliant to at least IEEE 802.11 a/b/g. Should support
Ad-Hoc and infrastructure modes.
d) Since the HTN protocol mandates that the roles of the nodes may change
depending on need, the nodes need to have some information stored on-board
which facilitates this role change. Hence, a storage medium like SD Card is
necessary.
e) The hardware must have capabilities to interface with a laptop or computer
easily for diagnostics.
48
f) Since this hardware will need to interface with other hardware and sensor
platforms, there must be multiple power supplies catering to various sensor
platforms.
g) There must be debug interfaces on the hardware for initial development
cycles.
h) The availability of development environment and APIs to reduce development
time.
i) Low bill of materials.
6.2. Block Diagram of Design
The following figure shows the basic building blocks of the design. In later
sub-sections the blocks are explained in more details.
Power
Supply
Microcontroller
K60RS-232
Clock
Circuitry
4 MHz
3.3 V
3.3 V
8-bit GPIO
Serial
Communication
LCDWi-FiZigbee USB SDHC
Sensor Board
Expansion
Connector
SPI at 4 MHz SPI at 12 MHz SDHC ModuleD+, D-, 5 V
3.3 V
3.3 V 3.3 V 3.3 V 3.3 V 3.3 V
3.3 V
Analog Lines & I2C
Figure 6.1: Hardware Block Diagram
49
6.3. Hardware Components
In the following each of the hardware components are introduced in detail.
6.3.1. Power Supply Circuitry
Figure A.1-4 in Appendix A shows the schematic of the power supply
circuitry. In keeping with the requirement of interfacing platforms with this hardware
that may need different levels of supply three different power supply options are
provided. The board has 5 V, 3.3 V and 1.8 V options. 3.3 V is the main power supply
voltage that runs all the peripherals on the board.
The LT1129-3.3 [47] IC is used to generate a 3.3 V from the 5 V supply. This
chip consumes only 50 uA while operating and can deliver up to 700 mA of current at
3.3 V.
TPS73701 [48] IC is used to generate 1.8 V from 5 V supply. This chip
consumes 0.3 mA while operating and can deliver up to 1 A of current at 1.8 V.
The supply of 5 V is taken from wall socket or USB and this is decided
through a selector switch. The two regulator ICs can be individually disabled.
6.3.2. Clock Circuitry
Figure A.5 in Appendix A shows the schematic of the clock circuitry. The
microcontroller runs off a 4 MHz external clock signal. The hardware board is given a
choice of sources for this 4 MHz clock signal. Using selector switches one can use a 4
MHz crystal, a 4 MHz clock chip or source the 4 MHz from an external clock source.
This is done to give maximum flexibility to the source of the clock signal for
the microcontroller.
The microcontroller also needs an external 32.768 kHz clock to operate.
50
An ECX-71 [49] tuning fork crystal is used for the 32.768 kHz clock source.
Its load capacitance is 12.5 pF.
An ECS-3951 [50] SMD clock oscillator is used as one of the two 4 MHz
clock sources for the microcontroller.
A Murata CSTCR4M00G53Z-R0 [51] ceramic resonator is used as the other
source for the 4 MHz clock for the microcontroller.
6.3.3. Microcontroller
Figure A.6-7 of Appendix A shows the schematic for the microcontroller.
The microcontroller used for this hardware is the Freescale Kinetis K60 [52].
The features of this microcontroller are presented in the list below:
a) Operates at a voltage of 3.3 V in this case.
b) 100 MHz ARM-Cortex-M4 delivering up to 1.25 Dhrystone MIPS per MHz.
c) 512 KB flash memory and 128 KB RAM.
d) 10 different low power modes of operation.
e) Built in security modules for integrity of firmware once burnt in the
microcontroller memory.
f) Two 16 bit SAR ADCs. Programmable Gain Amplifiers built in to the ADCs.
g) Two 12 bit DACs.
h) Three comparators containing a 6 bit DAC and a programmable input
reference.
i) Several timer modules available including real time clock.
The communication modules available with this microcontroller are as
follows:
51
a) Ethernet controller with MII and RMII interface to external PHY and
hardware IEEE 1588 capability.
b) USB full-/low-speed On-the-Go controller with on-chip transceiver.
c) Two Controller Area Network (CAN) modules.
d) Three SPI modules.
e) Two I2C modules.
f) Six UART modules.
g) Secure Digital host controller (SDHC).
h) I2S module.
As can be seen from above, this microcontroller fits the requirements
perfectly. The microcontroller has a maximum current consumption of 71 mA. The
SPI interface is used to connect to the Wi-Fi and ZigBee modules. The SPI maximum
speed is 12.5 MHz for the microcontroller. The abundance of analog lines on the
microcontroller makes interfacing external sensors easy. The presence of I2C bus
gives us the flexibility of interfacing digital I/O sensors as well. The integrated USB
transceiver makes interfacing this microcontroller to a laptop or computer very easy.
The microcontroller is driven at a core clock speed of 96 MHz. The bus clock
is 48 MHz. The USB module is driven at 48 MHz.
6.3.4. Wi-Fi Module
Figure A.8 in Appendix A shows the schematic of the Wi-Fi module.
The Wi-Fi module used in the design is an integrated solution with a
microcontroller and RF front end built on-module. The module interfaces to the
Kinetis microcontroller using the SPI bus at a SPI clock frequency of 12 MHz.
52
The Wi-Fi module used is the Redpine Signals module RS9110-N-11-22-05
[53]. Some of the capabilities of this module are listed below:
Operates at a low power supply voltage of 3.3V.
Small form factor of 22 mm X 28 mm
Integrates a uFL antenna connector for external antenna connection.
We use a 6 dBi gain rubber duck antenna.
Integrates full TCP/IP stack with an option of bypassing it in the SPI
mode. We use the TCP/IP stack within the module.
Fully compliant with 802.11 b/g and single stream 802.11 n standards.
Supports TCP, UDP, ARP, ICMP, IPv4 and DHCP.
Supports infrastructure, ad-hoc and power save mode of operation.
Current draw of 30 mA in transmit mode and 24 mA in receive mode
at 2 Mbps throughput.
Figure 6.2: General high level block diagram of the Redpine Signals Wi-Fi
module
53
The motivation behind using this module lies quite heavily on its low power
consumption, availability of low power modes to further reduce overall current
consumption and the integration of a fully functional Wi-Fi transceiver device with
minimal processing requirement for the host microcontroller.
6.3.5. ZigBee Module
Figure A.9 in Appendix A shows the schematic of the ZigBee module.
The ZigBee module used in the design is the Texas Instruments CC2530 [54].
The interfacing circuitry and RF front end around this chip is designed and
implemented. The RF front end consists of a signal conditioner, as the main
component. This is the TDK-DEA202450BT-7210A1 signal conditioning Band Pass
Filter for 2.4 GHz band. The circuit is connected to a SMA connector that uses a 6
dBi rubber duck antenna for transmission and reception of ZigBee signals.
Figure 6.3: Texas Instruments CC2530 System-on-Chip
54
The main characteristics of this chip which led to its selection are given
below:
Small 6 mm X 6 mm form factor with very few external components required
to setup functional ZigBee communication.
Standards compliant 802.15.4 implementation.
Low power consumption of 24 mA for transmit and 29 mA for receive.
Several low power modes are also available.
High sensitivity of down to -91 dBm.
High performance, low power consumption 8051 core with code pre-fetch
built in the module.
256 KB of FLASH which is enough to run the Z-stack ZNP firmware.
8 KB of RAM with retention in all power modes.
CSMA/CA hardware support.
Accurate digital RSSI/LQI value reporting.
The CC2530 is used in the ZigBee Network Processor configuration. The host
Kinetis microcontroller interfaces with this chip using SPI functioning at 4 MHz.
6.3.6. SDHC module
Figure A.10 Appendix A shows the schematic of the SDHC module.
The Secured Digital Host Controller [55] peripheral of the Kinetis K60
microcontroller is used to provide an on-board SD card storage option. In this
implementation a standard size SD card is used. The supply required for operation is
3.3 V.
The data communication can be made at a maximum of 25 MHz, but since the
requirements of this hardware do not place any specific constraint on this speed, for
55
the sake of data integrity a low speed mode of 400 kHz is used. The 4-wire mode is
used and hence at a time only a nibble can be exchanged physically from the
microcontroller to the SD card and vice-versa.
6.3.7. USB module
Figure A.11 Appendix A shows the schematic of the USB module.
The hardware is provided with the facility to connect to a laptop or computer
over USB. A micro USB connector is provided on the board for this purpose. The
USB transceiver is built-in the microcontroller. It implements USB according to the
standards specified at [56].
Special attention has been given to route the D+ and D- lines from the micro
USB connector to the corresponding pins on the microcontroller. Both the traces have
been kept the exact same length to prevent problems in USB communication that crop
up if the lengths are different.
The USB is also used to power the hardware. A selector switch is provided on
the power board to switch between wall-socket power supply and USB power supply.
6.3.8. RS-232 module
Figure A.12 Appendix A shows the schematic of the RS-232 module.
Although USB exists as a legitimate and high-speed mode for the hardware to
exchange information with a laptop or computer, but this mode is easy to implement
when only Mass Storage Device functionality is required. When program data needs
to be exchanged with the remote device, a driver and other requisite interfacing
software components need to be developed for proper communication.
Hence to lessen the development effort involved and also to keep alternate
modes of communication available the hardware is equipped with a DB9 connector
56
for communication with a laptop or PC using serial communication that can be
monitored on the PC side using readily available terminal software like
Hyperterminal. The RS-232 level conversion between the Laptop or PC and the
Kinetis microcontroller is done using the well-known Texas Instruments MAX3232
RS-232 transceiver IC [57]. The IC operates from the 3.3 V supply and draws a
maximum of 1 mA of supply current while operating. The transfer speed is a
maximum of 250 kbps.
6.3.9. LCD module
Figure A.13 Appendix A shows the schematic of the LCD connector.
The USB and RS-232 interfaces are good for connectivity with a laptop or PC.
When such a setup is absent, visual indication of processing state can be achieved
with the on-board display LEDs. But for many parameters and complex conditions
arising due to processing, LED display is not sufficient. Hence, the hardware is
provided with a LCD.
The LCD used is a 2.8 inch TFT display from Displaytech – SDT028ATFT
[58].
57
Figure 6.4: Displaytech SD028ATFT LCD with a picture displayed
It supports 240 RGB rows and 320 columns display. It has an integrated
display controller ILI9341 [56]. The LCD is equipped with backlight. It operates from
3.3 V power supply. The interface to the microcontroller is over an 8-bit parallel bus
made out of microcontroller GPIO lines and a few control lines also interfaced to
microcontroller GPIO lines. The connector from the LCD goes in to a 45 pin FPC
connector on the power board.
58
6.3.10. External Sensor Board Expansion Slot
Figure A.14 Appendix A shows the schematic of the sensor board expansion
slot.
The hardware is given the ability to interface external analog and digital
sensors or other relevant hardware through a 40-pin Hirose [59] connector. The
available analog input and output lines, not already used, from the microcontroller are
drawn to this connector. The I2C lines from the microcontroller are also drawn to this
connector so that sensors with digital output can also be interfaced to the hardware.
The need to supply power to the external sensor or hardware is addressed by tracing
power supply lines to the connector as well.
6.3.11. PCB Layout
In this section the motivation behind the mechanical design choices will be
presented shortly followed by the PCB layout of the hardware.
The number of components and peripherals on the hardware made the option
of going with a single board quite infeasible at the very beginning. That would have
resulted in a disproportionately large unmanageable board with long traces leading to
latency and increased power consumption.
Since the power supply to the board can be from varied sources like battery,
solar cells etc., it was decided to make the power supply circuitry separate from the
main hardware. There is a connector provided on the main board to which the power
board attaches. This gives the freedom to use different power boards, designed
differently with different components, with the main hardware as long as the pins on
the connector have the same lines connecting on both boards.
59
A large part of the power board was being left unused and the LCD had to be
on a visible part of the hardware. The placement of the LCD on the main board again
would have made the board very large. Instead the LCD was placed on the remaining
area of the power board with the signal lines being routed up from the main board
using board to board connectors. This not only results in efficient usage of space but
also lends physical stability to the board by providing two connectors to share the load
of the power board rather than just the one power connector.
Since both Wi-Fi and ZigBee have high speed signal lines going in to them,
they are placed on opposite planes of the main board. This prevents interference
between the high speed signal lines. Both the ZigBee and Wi-Fi portions of the circuit
are placed close to the edge of the board so that the antenna connectors are easily
accessible and also RF line trace lengths are short. The placement of the two parts of
the circuits are also as far apart from each other as possible in order to minimize
signal interference as they operate in the same frequency range although channel
assignments, power outputs and spreading characteristics are different.
The SD card slot and the sensor expansion board connector need to be reached
from the side as well as from the top, they are placed to the south end of the main
board. The top side reachability is also the reason why the power board and the main
board do not share the exact same dimensions.
To lend extra mechanical stability to the boards and provide easy access to the
underside of the hardware there are four nut and bolt arrangement with 5 mm spacers
in between holding the two boards together.
The PCB layout of the boards is provided in Appendix C.
60
6.3.12. HTNMote in Pictures
Figure 6.5: Power Board Top View
The above figure shows the Power Board from the top. The major components
on the board are marked in yellow.
The connectors (DB9 and wall socket power supply) are placed towards the
right hand top corner flush with the edge of the board.
The two switches seen in the figure are the power on/off switch and the switch
which selects whether USB or wall socket power supply is the source of power for the
board.
61
Figure 6.6: Power Board Bottom View
The above figure shows the Power board from the bottom. The two connectors
here are the board-to-board power connector (top center) and the board-to-board LCD
connector (left down). The power connector takes the power supply and RS-232
signal lines from the Power Board to the Main Board and vice-versa. The LCD
connector takes the signal and data lines related to the LCD from the Main Board to
the Power Board and vice-versa.
62
Figure 6.7: Main Board Top View
The above figure shows the Main Board from the top. The main components
on the board are indicated using yellow texts.
The mating connectors for the power and LCD connectors on the Power Board
can be seen.
63
There are some GPIO lines which we do not use on the microcontroller and
they are routed to the left side edge of the board so that we can use those lines
externally if required.
We can see that there is a guard band connected to the ground around the
clock circuitry near the microcontroller. This is necessary as the high frequency clock
sources are very sensitive to noise and external interference.
Figure 6.8: Main Board Bottom View
64
The above figure shows the Main Board bottom view. We can see that the
guards connected to the ground plane are again placed around the clock circuitry for
the ZigBee module. This is necessary as these high frequency clock sources are very
sensitive to noise and external interference.
The CC-Debugger connector is provided so that the internal workings of the
CC2530 chip can be traced and debugged during development.
Figure 6.9: HTNMote - Assembled
The figure above shows the HTNMote fully assembled. The Main Board and
the Power Board put together using nuts and bolts and spacers through the mounting
holes.
65
Figure 6.10: HTNMote – In action
The figure above shows the HTNMote powered by a wall socket supply. We
can see the green power LED indicator glowing. The LCD is programmed to display
some pictures and a line of text. The bank of indicator LEDs near the wall socket
power supply inlet is also glowing as they are connected to the data lines of the LCD.
66
Chapter 7. SOFTWARE IMPLEMENTATION
The development platform used for writing the software for this hardware is
the Freescale CodeWarrior for MCU v10.2 [60]. It is an Eclipse based IDE that allows
C/C++ to be used as the coding language of choice. It has a Processor Expert module
that slashes development time by automating low level register setup of the Kinetis
microcontroller. The user is presented with a GUI to select the properties that the
microcontroller needs to have for an implementation. The Processor Expert will
automatically generate the low level code required to setup the microcontroller. The
user can start from writing the actual system code by using the functions that the
Processor Expert exposes. All these functions are wrappers for low level system calls.
7.1. Wi-Fi Implementation
The Wi-Fi module from Redpine Signals comes with an API that can be used
to communicate over SPI with the host microcontroller [53].
This API contains two header files, namely rsi_config.h and rsi_global.h,
which consist of all the variable parameter values that can be set by the user. For
example, one such parameter can be the BSSID of the network to be created or joined
in ad-hoc mode.
The API communicates with the host microcontroller with a series of low level
hardware drivers that need to be modified according to the host microcontroller being
used. Here, these drivers have been modified to work with the Kinetis K60 SPI
peripheral.
The API consists of a central union that has all the variables needed to execute
commands on the module and also to receive responses from the module as a result of
67
executing those commands. The module carries the Wi-Fi functionalities out based on
the values that are found in this one central union.
The module expects a little Endian format from the host. The SPI signal
characteristics expected are a clock polarity of 0 and a rising edge clock phase. The
commands issued to the module consist of 4 bytes C1, C2, C3 and C4 which together
denote all the aspects of the command and ensuing data transfer to the module. Every
command in the API has a unique value for C1, C2, C3 and C4.
The API has six different types of commands namely, initialization, memory
read and write, frame read and write and register read.
The exchanges involved in each of these types are shown below [53]:
Figure 7.1: Slave initialization procedure
70
Figure 7.4: Memory Write or Master Write procedure
Figure 7.5: Frame write or Slave write procedure
71
Figure 7.6: Register read procedure
Utilizing this API the code for Wi-Fi send and receive functionalities has been
written. The corresponding flowchart is shown in Figure B.1 of Appendix B.
7.2. ZigBee Implementation
The ZigBee software has been implemented using the Z-Stack ZNP variant.
The CC2530 is setup to be a ZigBee Network Processor with the Kinetis K60 being
the host microcontroller communicating with the CC2530 over SPI. This scenario is
shown in the figure below:
72
Figure 7.7: CC2530 in ZNP configuration (Application Processor is K60)
The SPI is clocked at the maximum speed possible for CC2530, 4MHz. The
clock polarity is 0 and rising edge is used as the clock phase. MSB first bit ordering is
used. Standard 4 wire SPI is in use [54].
Apart from the standard SPI lines the CC2530 interface also requires two
additional control signal lines. These and their brief description are provided below
[54]:
a) SRDY: Slave ready signal. This signal is set by the CC2530 when it
is ready to receive or send data. It is set low when the CC2530 is
ready to receive data.
b) MRDY: Master ready signal. It is an active low signal. This signal is
set by K60 when it has data ready to send to the CC2530.
73
There are three types of command available under the CC2530 ZNP API.
These are AREQ, POLL and SREQ. The details of these commands can be found in
[54].
The CC2530 can be programmed as a ZNP in two modes, namely Simple API
mode and the AF-ZDO mode. In this research the AF-ZDO mode is used as it offers
better control over the CC2530 module as opposed to the Simple API.
The CC2530 ZNP software has three sections that it is divided into. These are
presented briefly below [54]:
a) SYS Interface: This provides the application processor a low level
interface to the CC2530 hardware and firmware. The peripherals on
the CC2530 accessible through this interface are the ADC, NV
items, hardware number generator and GPIO pins.
b) Configuration Interface: The commands in this interface allow us to
set up various parameters of the CC2530 radio chip.
c) AF-ZDO interface: This interface consists of the Application
Framework (AF) which is used by the application processor to
register its application with the CC2530 thereby being able to send
and receive data. The ZigBee Device Object (ZDO) interface allows
access to all network management related functions.
Utilizing API the software for ZigBee packet transmission and reception is
written. The flowchart outlining the code flow is shown in Figure B.2 of
Appendix B.
74
7.3. SDHC Implementation
The SDHC interface code is written using the Processor Expert exposed
function calls pertaining to the SDHC module in the K60. The block size of transfer is
512 bytes. A write, read-back and compare cycle is followed for every 512 bytes of
transfer to ensure integrity of data written on to the SD card.
The flowchart for the SD card interface code is shown in figure B.3 of
Appendix B.
7.4. USB Implementation
The USB on the hardware is implemented as a mass storage device with
vendor ID (VID) of 0x1234 and a product ID (PID) of 0x5678.
The flowchart for the USB implementation code is shown in figure B.4 of
Appendix B.
7.5. RS-232 Implementation
The serial communication software is set up to provide bi-directional
communication capabilities to ensure monitoring from the remote terminal side as
well as actuation from the remote terminal.
The software is currently setup to send a string for display on the remote
terminal. It then waits for 2 bytes of input to come from the remote terminal. Once the
inputs arrive they are echoed back for display on the remote terminal.
The communication baud rate is 9600 bauds and the character size is 8 bits
with no flow control.
The flowchart for the RS-232 implementation code is shown in figure B.5 of
Appendix B.
75
7.6. LCD Implementation
The LCD implementation is done using the embedded GUI (eGUI) [61] from
Freescale. This is a software suite that is used to interface any TFT-LCD with a
Freescale microcontroller.
This is a powerful but light-weight suite that has the capacity of rendering
colors as well as the following graphics objects:
Button
Check Box
Radio Button
Gauge
Icon
Label
Picture
Graph
Slider
Menu
Scroll Bar
Console
Text Box
It supports touch screen, multiple fonts and various screen sizes.
The structure of the software suites is layered. The lower layers contain the
low level drivers that interface the microcontroller with the controller on the LCD
module. As part of this research, new drivers were written to interface eGUI with the
ILI9341 [62] display controller chipset being used with the LCD.
76
The eGUI suite block diagram is show in the figure below [61]:
Figure 7.8: eGUI block diagram
The flowchart for LCD code is shown in figure B.6 of Appendix B.
77
Chapter 8. TESTS AND TEST RESULT ANALYSIS
Once the hardware platform was manufactured several tests were run on it.
Several of these tests were functionality tests to confirm whether the module in the
design was working as expected or not. The components mainly tested for
functionality only were USB, RS-232, LCD and SDHC.
The other components, namely Wi-Fi and ZigBee, were tested for
performance. This was essential as the HTN protocol needs the corresponding
hardware to meet certain performance benchmarks for it to be suitable for
implementation of the protocol.
In this chapter the first sub-section will consist of the test methodology, results
obtained and result analysis for the performance tests. The second sub-section will
consist of the functionality test outcomes.
8.1. Performance Tests
In the intended usage scenario of the HTNMote the most critical components
on the board, apart from the microcontroller, are the Wi-Fi and the ZigBee modules.
If these two modules do not perform above a certain level in comparison with other
off-the-shelf devices then the HTNMote solution will be rendered infeasible. The
following bunch of performance tests helps to show the capabilities of the
communication hardware on the HTNMote.
8.1.1. Methodology
The performance tests have been divided in to three groups, namely, the
current consumption tests, the Wi-Fi tests and the ZigBee tests. The methodology
followed for both these tests are described in the rest of this sub-section.
78
8.1.1.1. Current Consumption Test Setup
In this research work, the long-run steady state current consumption of various
peripherals on the HTNMote are measured. The test setup consists of 5 VDC, 1A
supply provided to the HTNMote using a wall power supply adapter. The voltage
regulators on the board convert this to the 3.3 V and 1.8 V on the board as required.
The current consumed from the wall socket source is measured. This gives us
an idea of the overall current consumption from all components and thus any battery
life model can be effectively come up with.
A multimeter is connected in series with the wall socket power adapter and the
HTNMote and the current readings are collected.
8.1.1.2. Channel Emulator
A wireless channel emulator from Azimuth Systems ACE 400WB [63] is used
to create wireless channels between the two end devices. The channel emulator is
shown in figure below:
Figure 8.1: ACE 400WB wireless channel emulator
79
This piece of equipment is a computer controlled device. It is able to create
accurate user-defined channel conditions between two devices at two ends of the
channel. This allows us to do accurate real-world tests without having to make
arrangements for elaborate field tests.
There are several advantages of using the channel emulator over real-world
field tests. They are listed below:
The use of the channel emulator allows the configuration of the
physical channel exactly according to the specifications of the user.
This means that an accurate physical channel containing a tree, a
house, two tall buildings and a stone can be emulated. Whereas,
finding a place in the field which has exactly this configuration of
physical structures is very difficult to find.
The other problem of the real-world is that the channel conditions are
never static and not repeatable. There are hundreds of parameters that
keep on changing every second in a real-world channel. Hence, if a
device needs to be tested in several configurations under equivalent
channel conditions then that becomes an impossible task in the field.
However, with the channel emulator the user has strict control over the
channel and hence repeatability is easily achieved.
A channel emulator is able to create any realistic real-world channel with great
accuracy. In this research work only the ITU-T standard channel Butler model [64] is
used. The power delay profile of the Butler channel model is shown in the figure
below captured from an actual active session of the channel emulator:
80
Figure 8.2: Power delay profile of the ITU-T Butler model
The Butler model is a static non-fading channel model that uses the identity
matrix for the channel coefficients. It does not have any multipath components or
scattering.
One of the common settings for the channel emulator common to both the Wi-
Fi and ZigBee tests is that the MIMO antennas were configured to have no
correlation.
8.1.1.3. Wi-Fi Tests Setup
The Wi-Fi tests consist of two tests that are performed to obtain the following
characteristics of the hardware:
a) Throughput vs. Packet size
b) Packet Loss vs. Path Loss
The channel emulator setup for the Throughput versus Packet size test is as
follows:
81
The path loss between sender and receiver fixed at 67.55 dB
The path loss between receiver and sender fixed at 67.65 dB
Input side attenuation Port A : 0 dB
Output side attenuation Port A : 0 dB
Input side attenuation Port B : 0 dB
Output side attenuation Port B side : 0 dB
WiFi channel used is channel number 1 centered around 2.412 GHz.
The channel emulator setup for the Packet Loss versus Path Loss test is as
follows:
Input side attenuation Port A : 0 dB
Input side attenuation Port B : 0 dB
WiFi channel used is channel number 1 centered around 2.412 GHz.
The layout and logical interconnection of the devices for the test is given
below:
Azimuth Director
Channel Emulator
HTNMote
Under Test
Laptop
Receiving Wi-Fi
Traffic
Figure 8.3: Layout and device interconnections for Wi-Fi tests
82
The channel configuration for the Wi-Fi tests on the channel emulator is
shown in the figure below:
Figure 8.4: The topology for the Wi-Fi tests
In the configuration above the device BS1 is connected to Port A1 of the
channel emulator. The BS1 is the HTNMote device under test. The MS1 is the laptop
which is the device that is the destination for the Wi-Fi traffic generated by the
HTNMote device under test.
The HTNMote device under test is configured to join an ad-hoc network with
SSID ‘FRAWSNNet’. It transmits at 16 dBm and the application code inside controls
the packet payload size for transmission.
83
8.1.1.4. ZigBee Tests Setup
The ZigBee module is tested for the following characteristics:
a) Throughput vs Path Loss
b) Goodput vs Path Loss
c) Packet Loss Rate vs Path Loss
The test setup includes the device under test which transmits ZigBee packets
of various payload size to the Coordinator device being monitored from Code Warrior
using the debugger.
A Texas Instruments ZigBee packet sniffer [65] is used to monitor all packet
exchanges happening in the channel. This allows us to later parse the packet trace and
find the throughput, goodput and packet loss rate parameters.
The layout and device interconnections for the ZigBee tests are shown in the
figure below:
Azimuth Director
Channel Emulator
HTNMote
Under Test
HTNMote
Receiver
CodeWarrior
Debug Session
Zigbee Packet Sniffer
Figure 8.5: Layout and device interconnections for ZigBee tests
84
The channel emulator is setup in the following layout:
Figure 8.6: ZigBee tests channel emulator topology
In the above layout the device under test is denoted by the BS1. This is a
HTNMote that is configured to be an end-device which sends ZigBee packets of
payload sizes 2 and 84 bytes to the coordinator device. A total of 10001 packets are
sent for every execution of a test.
MS1 denotes the coordinator device that receives packet transmissions from
the device under test. It is monitored via a debugger from the Code Warrior IDE.
MS2 denotes the ZigBee packet sniffer. It is kept parallel to the device under
test so that all the channel activity can be captured. The path from the device under
85
test to the sniffer is not attenuated. All attenuations are placed in the path from the
device under test to the coordinator device.
8.1.2. Current Consumption Test Results
The current consumption observed for various scenarios of interest are
tabulated below:
Scenario Current Consumption
HTNMote Switched on: voltage
regulators, power on LED, RS-232, SD
Card, USB, clock circuitry,
microcontroller active. No LCD.
48.4 mA
HTNMote Switched on and a single
indicator LED glowing
49.3 mA
HTNMote Switched on and RS-232
transaction on-going
51.5 mA
HTNMote Switched on and SD Card
transfer on-going
50.2 mA
HTNMote Switched on and connected
as USB mass storage device
59.3 mA
HTNMote Switched on and LCD
displaying graphic
77.5 mA
HTNMote Switched on and ZigBee
sending data to coordinator
80.1 mA
HTNMote Switched on and ZigBee 91.1 mA
86
receiving data from end-device
HTNMote Switched on and WiFi
module in idle but full power mode
163.4 mA
HTNMote Switched on and WiFi
module sending data
197.1 mA
HTNMote Switched on and WiFi
module sending data
190.1 mA
Table 8.1: HTNMote Current Consumption
When looking at the current consumption figures present in the table above we
must keep note of the fact that all of these figures are not stand-alone consumption
values for a corresponding component. Each of these values includes consumption
due to all peripherals active as well as the code running in the microcontroller. This
holistic view is necessary as in the field it is this overall current consumption which is
of importance rather than individual component consumption.
87
8.1.3. Wi-Fi Test Results
Figure 8.7: Throughput vs. Packet size test result
The above test is executed by keeping the channel conditions same and
varying the transmitted packet payload size from 50 bytes to a Wi-Fi maximum
packet payload size of 1400 bytes.
It is observed from Figure 8. That the lower packet payload sizes do not make
the device Wi-Fi operate at the saturation level of throughput and hence the
throughput gradually increases. As the packet payload size increases the throughput
starts to level off. It can be seen that using packet payload sizes of 800 bytes and
above will guarantee that the device is operating close to the maximum throughput
possible given all the other conditions.
It can be seen that the Wi-Fi module allows nearly 2.2 Mbps as the maximum
throughput.
0
500
1000
1500
2000
2500
50 100 200 300 600 700 800 900 1000 1100 1200 1400
Thro
ugh
pu
t in
Kb
ps
Packet Payload Size (in bytes)
WiFi - Throughput vs. Packet Size
88
Figure 8.8: Packet loss vs. Path loss result
The Figure above shows that the overall packet loss for the Wi-Fi module is in
the vicinity of a maximum of 32 packets per second when the payload size is 1400
bytes, the throughput is 2.2 Mbps and the path loss is 113 dB.
0
500
1000
1500
2000
2500
3000
68 73 78 83 88 93 98 103 108 113
Pac
ket
Loss
in
pac
kets
pe
r se
con
d
Path Loss in dBm
Packet Loss vs. Path Loss
50 Bytes
600 Bytes
1400 Bytes
89
Figure 8.9: Packet loss percentage vs. Path Loss result
The above figure shows the overall percentage packet loss observed for a
given path loss for different payload sizes at 2.412 GHz center frequency of
operation. It is of importance to note that the above result will be affected by change
in channel conditions and should be seen as the worst-case packet loss given similar
conditions. We can see that the overall packet loss increases dramatically as the
channel deteriorates. For smaller payload sizes we still are able to receive about 50%
of the packets when there is in excess of 100 dB path loss. But as the payload size
increases this amount becomes less and less and worst-case we observe near 100%
packet loss for 1400 bytes payload size at a path loss of 100 dB and above.
0
20
40
60
80
100
120
68 73 78 83 88 93 98 103 108 113
Pac
ket
Loss
in p
erc
en
tage
Path Loss in dB
Packet Loss Percentage vs. Path Loss
50 Bytes
600 Bytes
1400 Bytes
90
We use the following formula convert the path loss value into a distance
value:
(
)
In the above equation the path loss exponent (n) is assumed to be 2. The
distance (d) between the sender and the receiver is in unit of meters. The wavelength
(λ) is for the center frequency of 2.412 GHz. The path loss in the equation is in the
unit of decibels (dB).
Using the above formula we can see that for a distance of separation of
roughly 1 Km between the sender and the receiver the packet loss is a meager 10 - 12
packets per second when the transmission is very fast with 2.2 Mbps throughput using
1400 bytes packet size.
91
8.1.4. ZigBee Test Results
Figure 8.10: ZigBee throughput vs. path loss test result
The above figure shows that given the software interface provided by the
CC2530 Z-Stack ZNP API, the maximum throughput achievable is roughly 52 kbps.
We can see that this throughput value is towards the end of the graph where the
channel is substantially deteriorated. The increase in throughput comes from re-
transmissions in the channel.
0
10
20
30
40
50
60
45 55 65 75 80 81 82 83 84 85 86 87
Thro
ugh
pu
t in
kb
ps
Path Loss in dB
Throughput vs Path Loss
2 Bytes Payload
84 Bytes Payload
The throughput increases
when the channel
deteriorates due to the
increase in retransmissions.
92
Figure 8.11: ZigBee goodput vs. path loss test result
The term goodput denotes the number of useful information bits that were
transmitted per unit time from a source to the destination. As we can see from the
above figure the goodput falls as the channel deteriorates. This is obvious as packet
retransmissions and packet losses climb when the channel deteriorates. The above
figure also shows us that the current hardware will be able to receive transmissions
reliably down to 85 dB of path loss.
0
5
10
15
20
25
30
35
40
45 55 65 75 80 81 82 83 84 85 86 87 88 89 90 91
Go
od
pu
t in
kb
ps
Path Loss in dB
Goodput vs. Path Loss
2 Bytes Payload
84 Bytes Payload
93
Figure 8.12: ZigBee packet loss rate vs. path loss test results
The above result confirms what the test results for throughput and goodput
shows us. As the channel conditions deteriorate the number of retransmissions and
packet loss increases. This means there is more traffic being generated and
transmitted into the channel by the sender, this results in an increase in throughput.
But the number of transmissions actually reaching the intended recipient goes down
due to a bad channel and thus the goodput decreases.
0
50
100
150
200
250
300
45 55 65 75 80 81 82 83 84 85 86 87 88 89 90 91
Pac
ket
Loss
Rat
e
in p
acle
ts p
er
seco
nd
Path Loss in dB
Packet Loss Rate vs. Path Loss
2 Bytes Payload
84 Bytes Payload
94
8.1.5. Tests Result Analysis
In this section the analysis of the results obtained from the Wi-Fi and ZigBee
tests will be carried out to determine performance improvements that the HTNMote
offers over currently available sensor hardware platforms.
8.1.5.1. Wi-Fi Results Analysis
The Stargate gateway node [45] is a hardware platform that can interface with
the MICAz motes and provide a Wi-Fi gateway using a Wi-Fi Network Interface Card
in a PCMCIA slot provided in the hardware. The Wi-Fi transfer speeds on the
Stargate range from 1.5 Mbps to 3.75 Mbps depending on the Wi-Fi settings on both
the sender and receiver sides [66]. This is achieved at a cost of increased power
consumption as the Stargate operates at 5 V and consumes on an average 300 mA of
current when Wi-Fi is active [67].
A more recent development is the Wi-Fi expansion board [68] from Libelium
for the Waspmote sensor platform. This board interfaces to the microcontroller core
of the Waspmote using UART interface. This itself is a bottleneck for data transfer
between the microcontroller and the Wi-Fi radio. In a practical scenario, the current
consumed by the Waspmote hardware and the Wi-Fi board working together is close
to 200 mA [68]. The maximum transmission throughput achievable when the UART
baud rate is set to 57600 is 15.68 kbps.
The Wi-Fi results show us that there is a huge performance gain with
HTNMote when the parameter under consideration is the throughput. The HTNMote
provides a maximum throughput of close to 2.2 Mbps. This is an improvement of 140
times for similar current consumption.
95
Also when it comes to comparison with the higher power Stargate gateway, it
must be kept in mind that the maximum 1.75 times increase in throughput comes at a
significant increase in current consumption at a higher voltage, that is the power
consumption is much higher.
8.1.5.2. ZigBee Results Analysis
The goodput test results show that the HTNMote hardware is capable of
delivering up to 35 kbps in real-world application scenario at a path loss of 87 dB,
which is comparable to the performance of MICAz and Waspmotes.
This is expected since the radio chip used on the MICAz is CC2420 from
Texas Instruments and is a predecessor of the CC2530 used on the HTNMote but with
similar architecture. The XBee ZigBee module [69] used with the Waspmote is
theoretically capable of delivering better performance but is severely restricted by the
UART interface between the microcontroller and the XBee module.
8.2. Functionality Tests
The USB, RS-232, LCD and SDHC are components of the hardware whose
performance is not critical to the goals of the hardware but their proper functioning is.
The following demonstrate the result of running the software implemented to make
each of these modules work.
96
8.2.1. USB Test
The hardware is given a USB VID of 0x1234 and a PID of 0x5678. Upon
execution of the USB code, the device is plugged into a computer using a USB cable.
The device enumerates and the proper VID and PID can be seen from the Device
Manager console.
Figure 8.13: Device Manager screenshot for USB functionality test
97
8.2.2. RS-232 Test
When the software for RS-232 communication is executed, it is observed both
on the Hyperterminal screen on the computer and on the CodeWarrior console that bi-
directional communication is achieved successfully.
Figure 8.14: RS-232 bi-directional communication screenshot
98
8.2.3. LCD Test
The LCD successfully comes up after the code is executed as can be seen in
the figure below:
Figure 8.15: LCD output after code execution
99
8.2.4. SDHC Test
The SDHC software is executed to write a string to the SD card and then it is
read back and compared with what was written. The result of the execution is shown
in the figure below:
Figure 8.16: Partial screenshot of CodeWarrior console after SDHC code has
executed
100
Chapter 9. SUMMARY, CONCLUSION AND
FUTURE WORK
9.1. Summary and Conclusion
The Federal Railroad Administration is looking at WSN to monitor freight rail
in real-time or near real-time. The freight rail is an important contributor to the health
of the United States economy. Freight rail is also a greener mode of transport. The
current methods of monitoring freight rail, like wayside monitoring etc., suffer from
the drawbacks of not being able to monitor a large parameter set and the monitoring is
not in real-time.
WSN is a natural choice in achieving the real-time monitoring goals. ZigBee
is the protocol of choice for such deployments owing to its low cost and low power
consumption.
However, the railroads deployment scenario is very unique when it comes to
the topology in which the network exists. The long linear chain-like topology is one
for which ZigBee was not designed. Hence, the use of ZigBee in such scenarios
present significant problems.
These problems include:
a) Synchronization Delay
b) Route discovery issues
c) Packet and link loss
d) Lack of Quality-of-Service
e) Data forwarding and aggregation errors
f) Skewed network lifetime issues
101
A solution has been suggested by the Advanced Telecommunication
Engineering laboratory at the University of Nebraska-Lincoln which addresses the
above problems. This solution is called the Hybrid Technology Networking protocol.
This envisions ZigBee to operate in small clusters and the data forwarding from one
cluster to the next to happen using Wi-Fi. This has been shown to improve end-to-end
throughput, balance power consumption and also is standards based.
The current available sensor hardware platform does not have an integrated
solution for the application of Hybrid Technology Networking. Hence, the need to
design an integrated standards-based sensor hardware platform that will at least be
equivalent in performance.
The design of HTNMote is presented. It combines a Cortex M-4
microcontroller with both Wi-Fi and ZigBee radio capabilities residing on the same
board. SD Card storage facility is present. Communication with the external world is
carried out using RS-232 and USB apart from Wi-Fi and ZigBee.
The test results of the HTNMote hardware show that the Wi-Fi capability is
far better than any solution available in the market now. The ZigBee performance is
comparable to any other hardware currently available. Coupled with the possibilities
of implementing extremely low power consuming code, as all the peripherals support
power saving features, the HTNMote is an ideal platform to implement the Hybrid
Technology Networking protocol.
102
9.2. Future Work
The initial version of HTNMote has been shown to be quite capable of
handling the demands of the Hybrid Technology Networking protocol. It is already
better than comparable products available in the market.
In future further research on HTNMote will be conducted with the following
goals in mind:
a) Miniaturization: The next goal is to miniaturize the HTNMote hardware by
doing away with components that are not required in the field.
Miniaturization will further reduce latency and current consumption.
b) Improving performance: The current implementation has shown that even
though capable hardware and a fast interface to that hardware exists but
still performance bottlenecks come from the software implementation. In
the next revision, software optimization with an eye to throughput and
other performance improvement will be done.
c) Large deployment tests: In future, large deployments of HTNMotes will be
carried out to test how the HTNMotes perform when they are in a network
with several other HTNMotes.
103
REFERENCES
[1] http://www.fra.dot.gov/Page/P0362
[2] http://nationalatlas.gov/articles/transportation/a_freightrr.html
[3] AAR, Class I Railroad Statistics,
http://www.aar.org/PubCommon/Documents/AboutTheIndustry/Statistics.pdf
[5] http://www.railway-technology.com/contractors/track/trackside-intelligence
[6] http://en.wikipedia.org/wiki/Automatic_equipment_identification
[7] http://www.railway-technology.com/contractors/signal/transcore/transcore2.html
[8] http://en.wikipedia.org/wiki/IEEE_802.11
[9] Edwards, M.C.; Donelson, J., III; Zavis, W.M.; Prabhakaran, A.; Brabb, D.C.;
Jackson, A.S., "Improving freight rail safety with on-board monitoring and
control systems," Rail Conference, 2005. Proceedings of the 2005 ASME/IEEE
Joint , vol., no., pp.117,122, 16-18 March 2005
[10] http://www.gps.gov/
[11] I. F. Akyildiz, W. Su, Y. Sankarasubramaniam, and E. Cayirci. 2002. Wireless
sensor networks: a survey. Comput. Netw. 38, 4 (March 2002), 393-422.
[12] IEEE 802.11: Wireless LAN Medium Access Control (MAC) and Physical Layer
(PHY) Specifications. (2012 revision). IEEE-SA. 5 April 2012.
[13] http://www.wi-fi.org
[14] http://www.tutorial-reports.com/wireless/wlanwifi/wifi_architecture.php
[15] http://damayanthiherath.wordpress.com/a-survey-on-ieee-802-11-wireless-lan-
standards-and-physical-layer-issues
[16] http://it.siit.tu.ac.th/~u5322793290/its323/2.html
[17] http://www.radio-electronics.com/info/wireless/wi-fi/ieee-802-11-standards-
104
tutorial.php
[18] Lashkari, A.H.; Danesh, M.M.S.; Samadi, B., "A survey on wireless security
protocols (WEP, WPA and WPA2/802.11i)," Computer Science and Information
Technology, 2009. ICCSIT 2009. 2nd IEEE International Conference on , vol.,
no., pp.48,52, 8-11 Aug. 2009
[19] Majstor, F., "WLAN security threats & solutions," Local Computer Networks,
2003. LCN '03. Proceedings. 28th Annual IEEE International Conference on ,
vol., no., pp.650,, 20-24 Oct. 2003
[20] Sandirigama, M.; Idamekorala, R., "Security weaknesses of WEP protocol IEEE
802.11b and enhancing the security with dynamic keys," Science and
Technology for Humanity (TIC-STH), 2009 IEEE Toronto International
Conference , vol., no., pp.433,438, 26-27 Sept. 2009
[21] IEEE 802.15.4, “Part 15.4: Wireless Medium Access Control (MAC) and
Physical Layer (PHY) Specifications for Low Rate Wireless Personal Area
Networks (LR WPANs)”, IEEE SA Standards Board, 2003.
[22] http://www.radio-electronics.com/info/wireless/ieee-802-15-4/wireless-standard-
technology.php
[23] http://standards.ieee.org/about/get/802/802.15.html
[24] https://en.wikipedia.org/wiki/IEEE_802.15.4
[25] http://standards.ieee.org/about/get/802/802.2.html
[26] http://www.sensor-networks.org/index.php?page=0903503549
[27] Mulligan, Geoff, "The 6LoWPAN architecture", EmNets '07: Proceedings of the
4th workshop on Embedded networked sensors, ACM, 2007
[28] http://www.isa.org/
105
[29] David Flowers, Yifeng Yang, "The MiWi Wireless Networking Protocol is a
simple protocol designed for low data rate, short distance, low-cost networks.",
AN1066, Microchip.
[30] http://www.techonlineindia.com/article/12-02-
09/an_introduction_to_wireless_sensor_network_concepts.aspx
[31] http://www.radio-electronics.com/info/wireless/ieee-802-15-4/mesh-networking-
topology-topologies.php
[32] http://www.daintree.net/downloads/whitepapers/ZigBee_primer.pdf
[33] ZigBee Alliance: ZigBee Specification, ZigBee document 053474r13, 1
December 2006.
[34] Ata Elahi Ph.D.; Adam Gschwender,"ZigBee Wireless Sensor and Control
Network",ISBN-10: 0-13-713485-1,Publisher: Prentice Hal,October 29, 2009
[35] http://www.jennic.com/elearning/ZigBee/files/html/module2/module2-5.htm
[36] S. M. Rakshit, M. Hempel, H. Sharif, J. Punwani, M. Stewart, and S. Mehrvarzi,
“Challenges in current Wireless Sensor Technology for Railcar Status
Monitoring for North America's Freight Railroad Industry”, in
ASME/ASCE/IEEE Joint Rail Conference, 2012, pp. 1-9.
[37] J. Zheng, M. J. Lee, “Will IEEE 802.15.4 make ubiquitous networking a reality?
A discussion on a potential low power low bit rate standard,” IEEE
Communications Magazine, 2004, 42:140–146.
[38] P. Mahasukhon, H. Sharif, M. Hempel, T. Zhou, T.Ma, P. L. Shrestha, “A study
on energy efficient multi-tier multi-hop wireless sensor networks for
freight-train monitoring,” Wireless Communications and Mobile Computing
Conference (IWCMC), 2011 7th International,pp.297-301, 4-8 July 2011.
106
[39] P. Mahasukhon, H. Sharif, M. Hempel, T. Zhou, W. Wang, T.Ma, “Multi-Tier
Multi-Hop Routing in Large-Scale Wireless Sensor Networks for Freight Train
Monitoring,” ITS World Congress, 2010.
[40] P. Mahasukhon, H. Sharif, M. Hempel, T. Zhou, W. Wang, T. Ma, “Multi-Tier
Multi-Hop Routing in Large-Scale Wireless Sensor Networks for Real-Time
Monitoring,” IEEE Sensors, 2010.
[41] F. Osterlind, A. Dunkels, “Approaching the Maximum 802.15.4 Multi-hop
Throughput,” SICS Technical Report T2008:05, 2008.
[42] J. Misic, S. Shafi, V. B. Misic, “Avoiding the Bottlenecks in the MAC Layer in
802.15.4 Low-Rate WPAN”, in Proc. of 11th International Conference on
Parallel and Distributed Systems Workshops, 2005, pp. 363-367.
[43] S. M. Rakshit, M. Hempel, H. Sharif, J. Punwani, and M. Stewart, “Hybrid
Technology Networking: A Novel Wireless Networking Approach for Real-
Time Railcar Status Monitoring”, in ASME Rail Transportation Division Fall
Technical Conference, 2012, pp. 1-7.
[44] MICAz datasheet at MEMSIC:
http://www.memsic.com/support/documentation/wireless-sensor-
networks/category/7- datasheets.html?download=148%3Amicaz
[45] www.willow.co.uk/html/spb400-_stargate_gateway.html
[46] http://www.libelium.com/products/waspmote
[47] LT1129-3.3 datasheet: www.linear.com/docs/2235
[48] TPS73701 datasheet: www.ti.com/product/tps73701
[49] ECX-71 32.768 kHz crystal: http://www.ecsxtal.com/store/pdf/ecx-71.pdf
[50] 4 MHz SMD Quartz crystal: http://www.ecsxtal.com/store/pdf/ecx-32.pdf
107
[51] Murata 4 MHz ceramic resonator:
http://datasheetz.com/data/Crystals%20and%20Oscillators/Resonators/490-
1217-1-datasheetz.html
[52] Freescale Kinetis K60 microcontroller:
http://www.freescale.com/webapp/sps/site/taxonomy.jsp?code=
K60_ETHERNET_CRYPTO_MCU
[53] Redpine signals Wi-Fi module:
http://www.redpinesignals.us/Modules_&_M2M_systems/Modules/Wi-
Fi_Modules/Connect-io-n/RS9110-N-11-22.php
[54] CC2530: www.ti.com/product/cc2530
[55] SD Card Association: https://www.sdcard.org/
[56] USB standard: http://www.usb.org/developers/docs/
[57] RS-232 transceiver chip: www.ti.com/product/max3232
[58] Displaytech LCD: https://www.displaytech-us.com/2-8-inch-tft
[59] Board to Board Hirose connector datasheet:
http://www.mouser.com/ds/2/185/e53700036-7084.pdf
[60] Freescale CodeWarrior:
http://www.freescale.com/webapp/sps/site/homepage.jsp?code=CW_HOME
[61] Freescale eGUI:
http://www.freescale.com/webapp/sps/site/prod_summary.jsp?
code=EGUI&tid=vanEGUI
[62] ILI9341 display controller datasheet:
http://www.displayfuture.com/Display/datasheet/controller/ILI9341.pdf
[63] Azimuth Systems, http://www.azimuthsystems.com/platforms-channel-
108
400wb.htm
[64] http://www.itu.int/en/Pages/default.aspx
[65] TI ZigBee Packet Sniffer: http://www.ti.com/tool/cc2531emk
[66] http://platformx.sourceforge.net/Documents/nuts/WiFiFAQ.html
[67] Kumar, A.; Namboothiri, P.G.; Deshpande, S.; Vidhyadharan, S.; Sivalingam,
K.M.; Murty, S.A.V.S., "Testbed based throughput analysis in a Wireless
Sensor Network," Communications (NCC), 2012 National Conference on , vol.,
no., pp.1,5, 3-5 Feb. 2012
[68] Wi-Fi Expansion Board: http://www.libelium.com/expansion_radio_board/
[69] Xbee Module datasheet: www.digi.com/pdf/ds_xbeezbmodules.pdf
109
APPENDIX A
Please contact the author at [email protected] for more
information on this section.
APPENDIX B
Please contact the author at [email protected] for more
information on this section.
APPENDIX C
Please contact the author at [email protected] for more
information on this section.